Named after the hundred-eyed watchman of Greek myth, Argus watches the education landscape: spotting new opportunities, pressure-testing the ventures we're building, and tracing every read back to the real-world signals behind it.
The evidence library: the raw signals the pipeline is watching across the education ecosystem. Every idea is built from these.
arXiv:2608.14992v1 Announce Type: cross Abstract: Language-model systems increasingly read from stores they also write to, so a claim that was merely written earlier can return looking retrieved. We tested whether the message package carrying an unsupported assignment changes which answer a model gives in a synthetic lookup task. Claude Opus 5 selected a color code for a named item or abstained. In an exploratory four-arm study, false-code adoption was 0/24 with no target claim, 0/22 scorable trials when a prior assistant assertion named the target, 14/24 when a tool-result record named it, and 15/24 when that result used a ten-field metadata wrapper that marked it unchecked. The tool-result arm selected the record's code in 11/12 supported trials and 14/24 unsupported trials, ruling out a fixed output-token bias while leaving substantial planted-token heterogeneity. A document-preregistered replication reproduced the tool-result versus assistant-assertion gap, 7/24 against 0/24, one-s
arXiv:2608.14953v1 Announce Type: cross Abstract: Recent advances in Large Language Models (LLMs) have opened opportunities to apply high-level code transformations to the field of code optimization, and it has since emerged as one of the most fundamental tasks for LLMs to perform; however, at present, LLMs struggle to apply wide-ranging code optimization tasks due to both the complexity of the code and the inability to independently verify the correctness of the transformations. In this paper, we present the Trusted LLM (T-LLM) Compiler, which proposes an advancement in compiler technology through a collaborative effort involving high-level LLM code transformations, traditional compilers, and verification tools. Experimental results reveal that it can significantly improve code correctness when tested on a set of PolyBench/C benchmarks. Our approach facilitates iterative code optimization efforts with verification strategies that enable corrective actions. Through this approach, T-LLM
arXiv:2608.14945v1 Announce Type: cross Abstract: On-policy self-distillation (OPSD) gives language agents dense token-level supervision from a privileged self-teacher on the policy's own trajectories. Existing methods allocate this supervision mainly by teacher trust, but trust does not reveal whether emphasizing a token supports the current policy objective. We call this the trust-utility mismatch and introduce Influence Calibration for Self-Distillation (ICSD). For each supervised token, ICSD measures the first-order response of its importance-weighted RL surrogate contribution to a teacher-directed output perturbation. Batch-adaptive calibration converts this non-stationary signal into a bounded allocation weight while preserving the original auxiliary-loss mass within each action turn. These detached weights affect only the distillation loss and require no additional model pass. Across ALFWorld, WebShop, and Search-QA, ICSD improves all matched aggregate metrics over trust-only al
arXiv:2608.14944v1 Announce Type: cross Abstract: Natural-language interfaces can lower the barrier to programming robots, but existing systems struggle when users request complex tasks. While large language models (LLMs) perform well with simple commands, they often struggle to generate code for multi-step tasks, decompose high-level instructions, or reuse prior solutions. We present SkillComposer, an interactive natural-language robot programming system for simulation environments that continually learns reusable program abstractions. SkillComposer uses a generate-test architecture in which an LLM iteratively generates and revises robot programs before execution. Successful programs are stored and processed by an online library-learning algorithm that compresses recurring function sequences into reusable macro skills for future tasks. We evaluate SkillComposer through ablation experiments and a user study with 12 participants to determine its effectiveness on manipulation and robot c
arXiv:2608.14927v1 Announce Type: cross Abstract: Multi-agent large language model (LLM) systems can improve reasoning by spending more computation, but deployment requires deciding when extra collaboration is worth its cost. We isolate this decision by running every problem under four protocols while holding the solver fixed within each setting: direct solving (Baseline), iterative self-correction (Single), planner-executor-reviewer collaboration (PER), and multi-agent deliberation (Broadcast). The primary benchmark comprises 4,181 competition-level math problems; paired robustness checks cover four benchmarks spanning competition math, biology, and broader science with two solver families. Across fixed policies, trained routers, and frozen LLM routers, conservative policies under-escalate, whereas higher-solve frozen routers often over-escalate. A post-answer, pre-collaboration gpt-oss-120b probe ranks Baseline failures with 0.8847 AUROC (4,151 parseable cases; 95% CI [0.8732, 0.8955
arXiv:2608.14906v1 Announce Type: cross Abstract: Watermarking provides a principled way to authenticate text generated by large language models (LLMs). In practice, however, the final text may be mixed-source, with watermark evidence surviving at only a subset of token positions after rewriting, insertion, deletion, or paraphrasing. Although prior work has studied global detection of watermark signals, when such signals can be localized remains unclear. We formulate watermark localization as a token-level multiple-testing problem based on pivotal statistics, with a latent indicator recording whether watermark dependence survives at each position. Under an asymptotic regime indexed by exponents for signal sparsity, next-token concentration, and effective-vocabulary growth, we derive a sharp boundary for global detection and phase transitions for discovery and classification within the class of coordinatewise pivot-based localization rules. We show that discovery is strictly harder than
arXiv:2608.14881v1 Announce Type: cross Abstract: AI co-scientists that generate hypotheses, retrieve related work, design experiments, execute code, and draft full papers are beginning to change how research is carried out. Despite this rapid progress, state-of-the-art systems remain researcher-agnostic: given a research goal, they optimize novelty, validity, or reviewer score while ignoring the individual scientist who will use the output. This overlooks a fundamental fact about research, namely, that what counts as novel, valuable, or feasible depends on the researcher, including their prior work, methodological repertoire, and the collaborators and communities in which they are embedded. In this work, we introduce the problem of personalized auto-research, which conditions every stage of the research process on a representation of the individual researcher. We argue that personalization is not a convenience layer, but rather the fundamental property that allows an AI system to serv
arXiv:2608.14876v1 Announce Type: cross Abstract: Agentic coding assistants are finding widespread use, not just in new code development but in quickly ingesting and leveraging third-party code. This opens up a risk of malicious code being ingested as these coding tools operate with broad filesystem access inside developer workspaces. In this paper, we extensively study the impact of different dimensions of a novel attack surface we term workspace topology -- defined via directory depth, codebase modularity, in-file injection position and context framing -- on the attack success rate of adversarial prompt injection attempts. We perform an empirical study of indirect prompt injection (IPI) across a diverse set of open-source repositories spanning 10 languages and 6 engineering domains, evaluating three IPI entry points against open-weight models operating open source code harnesses. We find that workspace topology measurably affects IPI success. Specifically, changes in codebase modular
arXiv:2608.14838v1 Announce Type: cross Abstract: Retrieval components for code assistants are tuned against retrieval metrics: a configuration that raises recall@k is adopted, and downstream task success is assumed to follow. We report a controlled case study in code repair, not a new phenomenon but a deployed-flag, execution-graded instance of the known relevance-diversity and objective-mismatch tradeoff (Levy et al., 2025). On SWE-bench Verified we inject a retriever's hits as a fixed 12-slot context pack with no search tools and toggle one flag (one-chunk-per-file deduplication) on an otherwise identical stack. The flag is the higher-recall configuration (gold file present in 0.878 of served packs against 0.806 disabled), yet disabling it, trading file breadth for within-file depth, raises the single-shot resolve rate: gpt-5.6-sol +7.6pp (39.2% to 46.8%, n=500, McNemar exact p=0.0003), and a pre-registered open-weights replication any reviewer can re-run (Qwen3.6-27B, +3.6pp, n=499
arXiv:2608.14828v1 Announce Type: cross Abstract: Aligning a language agent to several objectives at once is a persistent failure mode of preference-based training: when objectives are combined additively, optimization collapses onto whichever is cheapest to improve and sacrifices the rest, so a support agent learns to sound warm while giving no real help. The root issue is that an additive reward has no notion of balance. We introduce Mint (MIN-selection preference disTillation), a one-line change to preference distillation: rather than ranking sampled candidates by a weighted sum of rewards, we rank them by their weakest objective, distilling the best-balanced candidate over the most lopsided one with an unchanged DPO objective. This is the p -> negative infinity limit of a generalized-mean family spanning additive to worst-case selection. Across cooperative emotional support and adversarial negotiation, min-selection lifts both objectives while sharply cutting their imbalance; on em
arXiv:2608.14808v1 Announce Type: cross Abstract: When a user question is underspecified, a capable model should recognize that its context is insufficient, identify the missing information, ask for it, and respond only once that information determines a unique answer. We formalize multi-turn information seeking as solving a k-underspecified constraint satisfaction problem, where k is the number of variables jointly required to determine the target and therefore measures the degree of missing information. We instantiate the formulation in MT-InfoSeek, a controlled evaluation suite of 5,251 problems and 9,006 task instances spanning mathematics, logic, biology, medicine, and general knowledge. We evaluate models along three axes: what they ask, when they ask it, and how the acquired information affects the final answer. Performance degrades across models and domains as underspecification increases. Models recognize that additional information is needed but underestimate how much, and in
arXiv:2608.14787v1 Announce Type: cross Abstract: Speculative decoding is a leading technique to reduce the cost of autoregressive generation by using a small drafter to propose several tokens, which are then verified in parallel by a larger target model. Speculative diffusion decoding (SDD) further removes sequential drafting by generating every position in a draft block in parallel with a discrete diffusion model. However, SDD still invokes the target on every block, leaving verification as a potential bottleneck. This paper recognizes that this creates a new control handle: whether to invoke the verifier at all. Thus, we study verifier skipping, a lossy policy that commits a selected draft prefix directly, and ask which confidence signal should schedule it. Interestingly, our study finds that better token predictors need not yield better schedulers: skips require contiguous high-confidence prefixes, while short skips can induce additional drafting rounds. To study this mismatch, we
arXiv:2608.14771v1 Announce Type: cross Abstract: Making language models solve constraint problems reliably often means having them translate the problem into a formal specification and delegating the search to a sound solver. But the translation is itself a language-model task, and an unfaithful translation makes the solver faithfully solve the wrong problem. Existing pipelines repair only translations that crash, returning the solver's error message and falling silent when the program runs but is wrong. We replace the error message with a proof: when the generated program is unsatisfiable, we extract a minimal unsatisfiable core over the model's own constraints and hand it back the exact set that cannot hold together, a leakage-free signal that localizes the fault. On a new benchmark of 77 problems with an exact oracle, translation to Answer Set Programming is faithful on six of seven domains and fails only on aggregate coverage scheduling, which concentrates the translation tax in o
arXiv:2608.14767v1 Announce Type: cross Abstract: Automated vehicles must explain their decisions in ways that passengers can understand, monitor, and trust. Existing language-annotated driving datasets are mostly observer-written, post-hoc, simulation-based, or generated from sensor inputs, rather than elicited from the driver performing the action. We introduce NARRATE, a multimodal real-world Australian driving dataset comprising 2,050 annotated events from 35 experienced drivers and driving instructors on public roads. Each event is grounded in synchronised visual, localisation, motion, and LiDAR streams and paired with in-vehicle and/or post-drive free-text explanations. NARRATE provides action labels, scenario-context labels spanning six high-level and 32 fine-grained categories, and span-level Situational Awareness (SA) annotations over driver explanations for Perception, Comprehension and Projection. Four benchmark tasks (SA, scenario-context, driver-action classification, and
arXiv:2608.14718v1 Announce Type: cross Abstract: Video understanding is a fundamental task for evaluating the capabilities of multimodal large language models (MLLMs). However, existing leading models have already achieved approximately 90% accuracy on the Video-MME leaderboard, suggesting that conventional single-turn video understanding tasks are becoming increasingly saturated and insufficient for assessing the intelligence of advanced MLLMs. Towards this end, we introduce VideoGAIA, an agentic video understanding benchmark for general artificial intelligence (AI) assistants. Moving beyond one-shot video question answering, VideoGAIA formulates video understanding as a multi-turn, tool-augmented interaction process, where models must iteratively perceive videos, invoke external tools, gather complementary information, and integrate multimodal evidence across turns. VideoGAIA contains 271 model-human co-designed tasks covering diverse and complex real-world scenarios. Each video-que
arXiv:2608.14710v1 Announce Type: cross Abstract: Predicting spatial gene expression from hematoxylin and eosin (H\&E)-stained images offers a cost-effective alternative to spatial transcriptomics (ST). However, existing methods treat H\&E images as generic visual inputs and ignore their intrinsic biological hierarchy, where spatially organized cell types collectively form functional tissue microenvironments that govern local gene expression programs. To bridge this gap, we formulate H\&E-to-ST prediction as a cross-modal semantic translation task and propose Path2ST, a hierarchically grounded autoregressive framework featuring three key components: (i) a Hierarchical Cell-Tissue Conditioning mechanism that fuses explicit and implicit cellular features with tissue-level semantic representations to construct hierarchical conditioning signals; (ii) a Scale-Adaptive Autoregressive Generation process over a hierarchical semantic vocabulary, enabling coarse-to-fine, biologically consistent
arXiv:2608.14658v1 Announce Type: cross Abstract: We introduce pico-type, a byte-level multi-head content classifier with approximately 1.5 million parameters that simultaneously predicts seven content properties from raw UTF-8 bytes in a single forward pass. Operating directly at the byte level -- no tokenizer, no subword vocabulary, no pretrained embeddings -- pico-type classifies coarse type (12 classes), modality (8), subtype (24), code language (62), text language (30), file MIME type (90), and risk flags (6-label multi-label: API keys, JWTs, passwords, emails, phone numbers, SSH keys). The architecture combines a learned byte embedding, three convolutional blocks with growing receptive fields, two bidirectional attention layers with rotary position encodings, and a statistical pooling layer feeding seven Matryoshka-style classification heads. Four tiered variants (tiny/small/base/pro) share the same trunk with sliced representations from 16 to 576 dimensions, yielding ONNX export
arXiv:2608.14644v1 Announce Type: cross Abstract: Real-world LLM deployments increasingly rely on runtime-injected prohibitions--enterprise policies, PII redlines, tool boundaries--that vary per request and per tenant. Conventional post-training is structurally ill-suited: SFT hides the violation signal in compliant labels, and DPO's sequence-level preferences mismatch token-localized violations. We propose DUET, a token-selective on-policy distillation method for prohibition compliance. DUET pairs a teacher that sees the prohibition (positive) with an identical-weight teacher that does not (negative). Because the two teachers differ only in prohibition visibility, their per-token disagreement isolates the prohibition's causal effect--yielding a clean supervision signal uncontaminated by model capacity or mismatch. This disagreement drives two complementary mechanisms: signal cleaning, which discards agreement tokens as redundant or prefix-corrupted, and preference-directed learning, w
arXiv:2608.14639v1 Announce Type: cross Abstract: Per-field accept/review with selective risk at most alpha -- accept a field only if the error rate among accepted fields is controlled -- is the trust contract document-extraction systems need, and the natural procedure silently violates it on real documents. On 13,859 genuine claude-sonnet-5 fields from 800 CORD receipts (49.0% correct) we diagnose three failure modes: document clustering (design effect 1.84-2.45), score-refit leakage (coverage 0.416 at risk 0.127, violating alpha=0.10 in 95% of splits), and a tie-mass pathology (a degenerate score collapses the threshold grid, 0.030 to 0.001). We organize the fixes as a validity ladder, guarantee form stated per tier. A fit/val split protocol restores expected-selective-risk control for a learned fusion: coverage 0.318 at risk 0.096 at nominal alpha=0.10, no tolerance band (production variant 0.326) -- an on-average point whose realized risk exceeds alpha in 47.5% of resplits, not a c
arXiv:2608.14617v1 Announce Type: cross Abstract: A recurring proposal in legal AI is to improve case-outcome prediction by fusing uncertainty tools (evidence graphs with belief propagation, sequential Bayesian odds updating, Dempster-Shafer combination, and conformal prediction) into one pipeline. We test this on 1,000 real European Court of Human Rights cases from LexGLUE and FairLex, predicting whether the Court found a Convention violation from the case's fact paragraphs. We compare three families across two frontier LLMs (Claude Opus 4.8 and GPT-5.5) as per-fact evidence estimators: (A) the raw LLM, (B) the LLM routed through the fusion pipeline, and (C) a term-frequency baseline through the same pipeline. Across roughly 4,750 tests we find: (1) on discrimination (AUROC around 0.83) the pipeline yields no improvement over either the raw LLM or the baseline; a frontier LLM used directly is the strongest single discriminator. (2) Naively composing an LLM with Bayesian-odds and Demps
arXiv:2608.14588v1 Announce Type: cross Abstract: Sequential multi-agent LLM pipelines chain specialized agents without verification at handoffs, creating a structural flaw with measurable and severe consequences. We show that hallucinations injected at Stage 1 do not merely persist; they transform: raw numerical facts become derived computations, then narrative prose, then editorially approved conclusions. At each transformation, detectability degrades near-irreversibly. We formalize this as the hallucination snowball effect, a first-order Markov process over four states (Raw Fact $\to$ Derived $\to$ Narrative $\to$ Invisible) with empirically measured per-boundary escape probabilities of 24.6%, 48.3%, and 89.3%. Across 346 automatically injected hallucinations in a 4-agent financial analysis pipeline on FinanceBench, gpt-4o detection drops from 72.0% at Stage 1 to 50.9% at Stage 4, and 23.7% of hallucinations survive completely undetected in the final output. Even the strongest model
arXiv:2507.15502v1 Announce Type: cross Abstract: Postoperative follow-up plays a crucial role in monitoring recovery and identifying complications. However, traditional approaches, typically involving bedside interviews and manual documentation, are time-consuming and labor-intensive. Although existing digital solutions, such as web questionnaires and intelligent automated calls, can alleviate the workload of nurses to a certain extent, they either deliver an inflexible scripted interaction or face private information leakage issues. To address these limitations, this paper introduces FollowUpBot, an LLM-powered edge-deployed robot for postoperative care and monitoring. It allows dynamic planning of optimal routes and uses edge-deployed LLMs to conduct adaptive and face-to-face conversations with patients through multiple interaction modes, ensuring data privacy. Moreover, FollowUpBot is capable of automatically generating structured postoperative follow-up reports for healthcare inst
arXiv:2608.16868v1 Announce Type: new Abstract: A language model's output does not by itself provide verifiable evidence about the internal computation that produced it. We study computational provenance: whether generated text can carry detectable evidence of which causally relevant internal state occurred. We test a bounded form of this idea in two controlled architectures: a modular feed-forward neural network and a transformer-based model. Both architectures are trained on the same arithmetic task with a mandatory pathway through two discrete intermediate states, allowing different internal paths to produce the same answer. We deliberately switch between these paths, authenticate the state actually used, and let that verified state determine a subtle statistical pattern in the generated text that can later be detected. The feed-forward and transformer systems each passed all 128 matched pairs in both their public and separately sealed protected end-to-end evaluations, with the dete
arXiv:2608.16834v1 Announce Type: new Abstract: We demonstrate that AI models are broadly susceptible to a phenomenon we call model hypnosis, in which individually weak and seemingly irrelevant cues in the prompt can be systematically combined to strongly control model behavior. Model hypnosis occurs across model families and scales, including in frontier reasoning models, and hypnotic prompts can transfer between models. Because the model is controlled by inconspicuous textual choices, such as paraphrases and typos, model hypnosis presents new challenges and avenues for AI safety, and is a major hurdle for AI interpretability.
arXiv:2608.16798v1 Announce Type: new Abstract: Agent harnesses have substantially improved performance on long-horizon tasks by coordinating agent interactions with the environment. However, reinforcement learning through complex harnesses remains largely unexplored, as scaling such training to long-horizon agent tasks introduces fundamental challenges. In this work, we present a unified black-box RL framework for stable and scalable optimization of general agents through complex harnesses. Concretely, we first build a sandbox-based execution infrastructure that isolates task environments and harnesses within temporary sandboxes for large-scale concurrent rollouts. We then decouple policy optimization from opaque harness execution and place a serving proxy at the model boundary to capture model calls. To reconstruct multi-turn trajectories and improve training efficiency, we organize the captured calls into prefix trees and further adapt both critic-based PPO and critic-free GRPO to o
arXiv:2608.16707v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly deployed as decision-making agents in settings that require sophisticated environmental exploration. However, existing work has raised questions about how LLMs actually balance exploration and exploitation. Unlike classical agents, LLM agents engage with tasks through natural language, exposing them to semantic information with no formal counterpart in the task structure. We introduce the semantic bandit, an extension of the multi-armed bandit setting that explicitly considers the textual labels assigned to actions, and use it to study how semantic priors --- inductive biases arising from associations between language and expected reward learned during pre-training, shape LLM exploration behaviour. We find that semantically informative action labels reduce exploration in favour of exploitation, improving performance when aligned with the reward structure and severely degrading it when misalign
arXiv:2608.16671v1 Announce Type: new Abstract: The language-model head maps a hidden state of width D to a vocabulary of size V, so its transpose can return at most D independent directions to the Transformer. Godey and Artzi argue that this severe projection is a harmful optimization bottleneck. We separate the geometry from the causal claim. Our backward-only intervention keeps the ordinary logits and the exact LM-head parameter update while reducing only the rank of the gradient sent into the Transformer. Across five paired seeds on byte-level and BPE-8192 WikiText-2 models, reducing backward rank increases validation loss. An equally ranked factorized forward head, however, increases loss substantially more. At half rank in the larger model, the backward-only loss increase is 0.0586 (95% CI [0.0167, 0.1005]), while the factorized forward head increases loss by 0.1795 ([0.1547, 0.2042]). The vocabulary-space residual also contributes to the ordinary LM-head update, and removing tha
arXiv:2608.16650v1 Announce Type: new Abstract: Large language models (LLMs) exhibit sycophancy, a tendency to agree with user beliefs regardless of factual accuracy. This can reinforce misconceptions, but eliminating it entirely risks over-correction against valid opinions. Effective control must therefore both reduce and increase sycophancy with predictable and gradual effect. Yet, existing methods fail to ensure a bidirectional and monotonic relationship between steering strength and behavioral outcome across models and datasets. We introduce PCA-guided Activation Scaling (PAS), an activation steering framework that decomposes residual stream activations into a PCA-identified sycophancy-honesty subspace and an orthogonal residual, then applies distinct scaling exponents to achieve monotonic, bidirectional control. Across three LLMs and three datasets, PAS achieves strong monotonicity (Spearman $\rho$ = +0.92) and an average shift of 15.4% per direction, compared with 8.7% for the ba
arXiv:2608.16647v1 Announce Type: new Abstract: On-policy distillation (OPD) transfers teacher capabilities by supervising trajectories sampled from the student's own policy, yet its generalization behavior remains poorly understood, as most studies evaluate OPD on a single domain and on benchmarks close to the training data. We present a controlled study that varies one generalization factor at a time, from in-domain distribution shifts to cross-domain transfer and the multi-teacher setting. We find that OPD transfers a teacher's reasoning behavior rather than its answers to particular problems: training difficulty barely matters, and even problems the teacher never solves are useful. Transfer depends strongly on the origin relationship between teacher and student: same-origin pairs bring the student close to the teacher across languages, reasoning horizons, and even other domains, whereas cross-origin pairs mostly fit the trained distribution. This broad reach is a double-edged sword
arXiv:2608.16643v1 Announce Type: new Abstract: Automated detection of errors in clinical documentation is a promising application of large language models (LLMs), yet decisions to deploy such models rest on benchmarks that evaluate each clinical note in isolation. Error-detection benchmarks are typically constructed by injecting errors into notes, such that each erroneous note has a natural counterpart. Aggregate discriminative metrics (e.g., balanced accuracy or F1) do not exploit this structure. We show that this omission is consequential. In particular, evaluating 15 diverse LLMs on 4 standardized clinical error-detection test sets across 3 languages, we find that 13 of 15 models fall below the level of random pairwise discrimination, even while achieving F1 scores that standard practice would read as moderate. We also observe that the underlying bias patterns differ across languages: the same model can default to "no error" on one language and over-flag errors on another. To diagn
arXiv:2608.16627v1 Announce Type: new Abstract: Natural language explanations (NLEs) are increasingly used as inputs, for example, as few-shot rationales that influence model behavior in in-context learning (ICL). However, it remains unclear how different types of NLEs compare in their effects on downstream model performance in explanation-augmented prompting. Therefore, we provide a comparative evaluation across six benchmarks and four instruction-tuned models, studying how NLE source (human-written when available, self-generated explanations, generated by an external LLM) and NLE selection (random vs faithfulness-based filtering) affect downstream utility of NLEs when used in ICL settings. Our extensive evaluation shows that, on classification-style benchmarks, adding NLEs to few-shot prompts often improves accuracy over few-shot prompting without explanations; among NLE sources, externally generated LLM-NLEs often provide strong downstream utility and remain competitive with human r
arXiv:2608.16620v1 Announce Type: new Abstract: Palmyra x6 is a large language model optimized for use with enterprise-oriented agentic tasks. The model was built by post-training a Mixture-of-Experts base model with Anchored Supervised Fine-Tuning on a compact corpus of verified, synthetic tool-use trajectories, optimized with a Muon + Adam hybrid. The recipe is deliberately conservative and deliberately controlled: 626 trajectories, a single epoch, a low learning rate, and a KL anchor to the frozen base. The model shows substantial gains over the previous default model for Writer Agent, and compares favorably with several recent models on public benchmarks, scoring the highest on BFCL Core at $0.785$ and posts the highest six-benchmark mean of the cohort. Furthermore, the model has shown itself to be competitive or leading relative to comparators in our bias and safety evaluations.
arXiv:2608.16577v1 Announce Type: new Abstract: Large language models (LLMs) are deployed globally in high-stakes settings, yet most safety research and alignment efforts remain concentrated on English. Thus, users interacting with LLMs in other languages may encounter weaker safeguards despite relying on the same systems for similarly sensitive tasks. In this work, we investigate whether safety signals learned from a high-resource language, like English, can improve multilingual safety. We propose BabelSteering, an activation steering method that acts as a lightweight inference- time intervention, using refusal directions derived from English safety supervision to generalize across languages. Our evaluation includes eight languages and jointly measures refusal of harmful requests, over-refusal, and general task utility. The results show that BabelSteering increases the refusal of harmful requests across languages, with only a marginal to no reduction in task utility but with some incr
arXiv:2608.16554v1 Announce Type: new Abstract: Answer-only reinforcement learning (RL) trains reasoning models to solve fully specified problems, but many realistic queries omit a premise needed for a unique answer. In this setting, the useful response is not always refusal: the model should ask for the missing premise, condition its answer on the unknown quantity, or abstain when no informative conditional response is available. We present \emph{Ask-Condition-Abstain Reinforcement Learning} (ACA-RL), a data-augmented RL framework for this setting. Its reasoning-graph-guided pipeline converts well-posed problems into missing-premise training instances with localized gap annotations; ACA-RL then trains on these instances with a structured reward over five observable response behaviors. We also introduce the \emph{Missing-Premise Benchmark} (MPB), a 274-instance human-verified benchmark spanning mathematical, logical, and real-world word problems. Across Qwen3 and Llama models, ACA-RL c
arXiv:2608.16553v1 Announce Type: new Abstract: Multi-preference alignment is often framed as scalarization: combine reward dimensions, then optimize. This leaves a temporal decision underspecified: when should each preference dimension enter policy optimization? We propose \methodname, a stability-guided active-set controller for controlled objective admission. \methodname starts from a small active set, retains admitted objectives, and expands when reward-deviation gates indicate low recent deviation or a patience budget is exhausted. A probing phase estimates a hard-to-easy order, and adaptive weighting emphasizes underperforming active dimensions. Automatic evaluations with 15 training preferences and 16 held-out benchmark columns show that \methodname obtains higher averages than simultaneous scalarization and shared-budget adapted baselines. Component ablations and expansion dynamics further support cumulative retention, gated admission, and probing-derived ordering as useful des
arXiv:2608.16515v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) improves large language models by grounding generation in external evidence, but it also introduces a source trust problem: retrieved context may be useful, irrelevant, or even misleading. Existing RAG systems often apply a fixed trust policy toward retrieved evidence, which can either over-trust incorrect context or underuse context when the user explicitly asks for context-following behavior. Therefore, we propose Intent-Guided Decoding (IGD), a framework that arbitrates between retrieved context and parametric memory according to user intent. IGD uses answer-level filtering and token-level correction to steer the final decoding trajectory between retrieved context and parametric memory. We evaluate IGD on three faithful QA benchmarks and three factual-conflict benchmarks across five LLMs, IGD substantially improves factual recovery, achieving gains of up to 65.4 percentage points on factual-conflict
arXiv:2608.16417v1 Announce Type: new Abstract: Multi-modal retrieval-augmented generation (RAG) is a key technique for visually rich long document understanding. Existing multi-modal RAG methods are progressively advancing toward multi-agent systems: they first retrieve relevant pages based on a query, and then iteratively understand information within those pages. However, these methods typically rely on fixed workflows and lack the ability to dynamically scale computation at test time, often leading to insufficient evidence. To address this, we propose D2-ScaleAgent, an agentic framework that introduces a dual-dimensional scaling paradigm for retrieval and reasoning. The core of D2-ScaleAgent is a Verifier agent-driven dynamic routing loop based on the intrinsic difficulty of the query, centered around a continuously updated evidence bank that serves as the agent's dynamic working memory: when retrieval needs to be expanded, the agent routes outward (retrieval scaling), decomposing
arXiv:2608.16390v1 Announce Type: new Abstract: PDF corpora advertise their size in tokens but compute every rate they publish (coverage, OCR routing, re-fetch recovery, language mix) per document, and none decomposes its token total. The two units diverge sharply. On CC-MAIN-2021-31-PDF-UNTRUNCATED (7.9M web PDFs, 32.6B tokens), 3.02% of text-bearing documents hold half the tokens (Gini 0.807); documents over 50 pages are 5.00% of the corpus but 53.53% of its text. The PDFs produced by a TeX{} toolchain are 1.66% of documents and 4.05% of the text. The clearest casualty is Common Crawl's truncation cap: it affected 23.06% of documents and 63.08% of the text. Reconstructing the truncated files and extracting both versions, two widely used libraries recover 11.4% and 1.4% of that text; between 72% and 97% of affected documents yield nothing; roughly 55--62% of the corpus's text is lost. Under the 5 MiB cap adopted in March 2025, 30.19% of tokens would still be truncated, and recovery on
arXiv:2608.16386v1 Announce Type: new Abstract: Financial agents must do more than recall domain knowledge: they must be both reliable, executing precise operations over grounded evidence, and executive, sustaining long-horizon research whose conclusions remain auditable. We present Mint-Agent, a family of finance-native agentic models designed around these two scales of financial intelligence. Mint-Agent is built upon three pillars: data, harness, and algorithm. Our data engine constructs clean, specialized tasks for atomic financial capabilities and long-horizon agentic execution from real-world financial sources. MintHarness enables stable interaction with open-ended environments and maintains auditable evidence trails across extended research trajectories. Our training recipe combines SFT, critical-step OPD, and RLVR to develop separate financial reasoning and agentic execution experts, which are then unified through model merging and multi-teacher on-policy distillation into compa
arXiv:2608.16379v1 Announce Type: new Abstract: Evaluating speech recognition for a Kurdish variety written in a Latin field orthography, using a model that outputs Arabic script, creates a measurement problem before a modelling one: direct scoring treats writing-system differences as recognition errors. Jointly normalizing reference and hypothesis avoids this, but also changes reference tokenization, mixing agreement gains with a change in the scoring denominator. I evaluate MMS-1B-all with the Central Kurdish (ckb) adapter, used as released without adaptation, on 1,722 Garrusi questionnaire segments from five speakers (9,763 reference word tokens; 117.9 minutes). I use a common-reference design: the reference is folded once and fixed at 9,763 tokens, while only the hypothesis representation varies. The raw Arabic-script hypothesis scores 111.70% WER and 100.92% CER, with zero exact word matches. Latin transliteration gives 102.36% WER and 57.89% CER; folding it into the reference's r
arXiv:2608.16353v1 Announce Type: new Abstract: Even well-aligned large language models confidently generate factually incorrect text, making hallucination a persistent reliability risk in high-stakes deployments. These models nonetheless carry linearly separable truthfulness signals in their internal representations. Existing white-box detectors, however, collapse this evidence to isolated components or a single depth, discarding discriminative information distributed across the full forward pass. We introduce HalluTracer, a detection framework that reads and aggregates truthfulness evidence across every layer of the forward pass before the model emits any answer token. A geometric analysis reveals that the per-layer signals are weakly correlated, so that simple depth averaging suppresses layer-specific noise and captures nearly all linearly accessible information. Across six open-source language models and five hallucination benchmarks, HalluTracer consistently outperforms matched wh
arXiv:2608.16347v1 Announce Type: new Abstract: Direct communication between AI systems relies on natural language as an intermediate layer, incurring encoding/decoding overhead, token cost, and latency. We ask whether internal activation states can instead be transferred causally between different large language model (LLM) architectures via a learned projection, evaluated at three levels: representational similarity, cross-model retrieval from projected states, and end-to-end causal transfer via activation injection during generation. Using four architecturally diverse open-weight models (Qwen2-0.5B, Phi-3-mini, Mistral-7B, FLAN-T5-base), we find that representational alignment in trained models exceeds a random-initialization null baseline and is best captured by a rank-based metric (mutual k-nearest-neighbour alignment), more robust to activation-magnitude outliers than centered kernel alignment (CKA) or Procrustes analysis. A learned projection network retrieves the correct target
arXiv:2608.16344v1 Announce Type: new Abstract: Indic quality estimation (QE) and automatic post-editing (APE) data is spread across separate releases, so no single resource supports training and evaluation across tasks and language pairs on one footing. We consolidate the WMT 2020--2024 shared-task lineage with an extended English--Malayalam resource into \indicqe: $126{,}754$ instances over nine directional pairs, with up to four label types aligned on the same segment, a direct assessment, a human post-edit, word-level OK/BAD tags and an error explanation, and a test set stratified over four difficulty axes. On it, we benchmark six prompted LLMs and three COMET metrics on segment-level QE, and three systems on APE. Two of the axes are defined partly on the direct assessment and select a compressed slice of it, so each axis is compared against a control drawn from the same language pair with the same score distribution. Only one survives that control: segments whose holistic and toke
arXiv:2608.16333v1 Announce Type: new Abstract: On-policy distillation (OPD) aligns a student model with a teacher's logit distribution on student-generated trajectories. This approach has achieved strong empirical gains and can often surpass conventional off-policy distillation with substantially less data. However, standard token-level OPD can provide only fragmented corrections along an erroneous student trajectory and cannot unfold a complete and correct repair path. Motivated by this limitation, we propose \emph{Step-Level On-Policy Distillation} (SOPD), which combines the long-horizon correction of supervised fine-tuning (SFT) with the on-policy advantage of OPD to provide step-level supervision over complete student-generated trajectories. We show that, at different limits of step length, SOPD reduces to SFT or approximates OPD. Compared with SFT, the teacher responses in SOPD are conditioned on student trajectories and therefore align more closely with student-visited states; c
arXiv:2608.16303v1 Announce Type: new Abstract: Long-term emotional-support agents require memory mechanisms for personalized understanding across sessions. However, emotional-support dialogue is often low-density: turns are incomplete, evidence is scattered, and user states evolve over time. Existing memory methods usually rely on fixed units, such as turn-level notes or session summaries, which may lose details or introduce redundant noise. We propose FTA-Mem, a structured memory framework for low-density long-term dialogue. FTA-Mem uses Boundary-preserving Window Segmentation (BWS) to form coherent situation fragments, and constructs Fact-Time-Affect Memory Units (FTA Units) that jointly encode factual content, temporal grounding, and affective context. Retrieved units are then synthesized into structured context for answer generation. Experiments on ES-MemEval and LoCoMo show that FTA-Mem improves overall long-term memory question answering across benchmarks with different informat
arXiv:2608.16295v1 Announce Type: new Abstract: AI coding agents need more than relevant snippets: they need business semantics, validation evidence, relations, and assurance that their context is current. Existing systems usually infer or externalize this knowledge through retrieval, summaries, graphs, rules, or reverse specifications. We investigate a complementary representation in which selected code units directly carry agent-usable knowledge. We introduce Executable Code Knowledge (ECK) and define an Executable Code Knowledge Unit (ECKU) as a source-bound object combining stable identity, semantics, executable behavior, contracts, evidence, relations, provenance, validation state, and a query interface. Our Python prototype supports code-local authoring, manifest export, evidence execution, exact changed-line impact, freshness checking, and agent-facing projections. Across three real Python repositories and 26 controlled patch tasks, direct ECK provides executable test coverage f
arXiv:2608.16286v1 Announce Type: new Abstract: While various organizations now actively encourage LLM use in classrooms, we still lack rigorous, systematic evaluations of how well these models actually perform the fundamental tasks of language pedagogy. This paper examines whether state-of-the-art LLMs can deliver the kind of corrective feedback and methodological explanations that language learners need. The study tests multiple large language models on their ability to identify, correct, and explain common learner mistakes in English, by systematically varying model parameters to investigate how these technical adjustments affect output quality, pedagogical clarity, and consistency, along with using retrieval-augmented generation to query methodological data. The evaluation employs automated metrics (GLEU, BERTScore) but also human expert judgments to capture dimensions that purely computational measures miss: linguistic nuance, cultural sensitivity, and instructional appropriatenes
arXiv:2608.16269v1 Announce Type: new Abstract: Recent advances in neural topic models with pre-trained language models (PLMs) have achieved strong performance by leveraging general-domain pre-training, yet their topic interpretability often degrades on specialized corpora. This limitation primarily stems from the geometry of the embedding space, where domain-specific terms unseen during pre-training collapse into an indistinguishable region, and neither domain-specific re-training, word-level graph enrichment, nor parameter-efficient fine-tuning can restructure this space without inheriting the capacity ceiling of the underlying encoder. Our key insight is that a learnable graph layer operating on token-level PLM embeddings can acquire corpus-specific semantic structure that the frozen encoder lacks, because token-level graphs preserve document-local context that word-level representations discard and joint optimization with the topic objective reshapes embedding geometry directly fro
arXiv:2608.16224v1 Announce Type: new Abstract: By leveraging large-scale pretraining, LLMs can interpret diverse temporal expressions and question formulations without task-specific training. However, existing prompt-based neuro-symbolic systems continue to rely on LLMs for both semantic interpretation and exact temporal inference. Consequently, discrete decisions regarding intervals, time anchors, and ordered states remain vulnerable to probabilistic errors and difficult to verify. We present STAIR, a \textbf{S}emantic-\textbf{T}emporal \textbf{A}utomaton for \textbf{I}nterpretable \textbf{R}easoning. STAIR separates semantic interpretation from precise temporal inference: an answer-free LLM adapter maps complex question formulations to normalized temporal intents, while a deterministic temporal automaton with finite control and guarded transitions executes the corresponding policies over canonicalized evidence. Following a rule-first design, STAIR resolves standard questions without
arXiv:2608.16185v1 Announce Type: new Abstract: LLM agents increasingly answer questions over dynamic raw-document collections, where files may change before preprocessing, and relevant evidence (spans, sections, pages, or tables) is query-dependent. Existing retrieval-augmented approaches pre-materialize evidence via fixed chunking, embeddings, or persistent indexes: effective for lookup, yet costly, stale-prone, and committed to a granularity before the query is known. We formulate in-context search as Budgeted Evidence Localization over a latent evidence space induced by dynamic raw documents and propose LENS (Latent Evidence Exploration and Search), an index-free framework. Instead of pre-materializing the evidence space, LENS maintains a query-conditioned belief over candidate units, iteratively selecting candidates via complementary lexical, local, and exploratory proposal policies, updating the belief via an LLM relevance oracle, and narrowing toward high-posterior regions under