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.
The county education office cited the cost of recently approved employee contracts as well as ongoing structural deficits and declining enrollment. The post LAUSD placed on increased fiscal oversight with warning of looming financial troubles appeared first on District Administration .
For years, the conversation around young children and screens has been dominated by a fear of too much time, too little interaction and too many missed opportunities for real learning. In many cases, those concerns are justified. After all, research consistently shows that children’s excessive or passive screen use, especially of entertainment-heavy content, can negatively […]
Schools today are navigating an unprecedented convergence of academic gaps, behavioral challenges, chronic absenteeism and rising student mental health needs.
When Jon Ketler, founder of the Tacoma School of the Arts, passed away recently, Getting Smart contributor Chris Unger sat down to capture what made him extraordinary. This tribute is more than a remembrance. It is a masterclass in what visionary school design actually looks like when one person decides to stop waiting for permission and start building. Education leaders who have ever wondered whether a single individual can change a system will find both proof and inspiration here. The post He Showed Me What One Person Could Do: A Tribute to Jon Ketler appeared first on Getting Smart .
The continuing popularity of geolocation challenges suggests that people have not lost their appetite for discovery–and critical thinking.
Centralized application services transformed how students apply to medical school. They gave applicants a single point of entry, standardized how programs receive and review data, and provided comparable information institutions could actually use. The post The case for centralized admissions in graduate medical education appeared first on eCampus News .
The Hidden Partner in Workforce Development Elizabeth Redden Fri, 07/10/2026 - 03:00 AM Federal policy is focused almost exclusively on entry-level workers—and it’s leaving behind college-educated professionals in need of reskilling. Byline(s) Hugo Villar
Global Cooperation Needed to Tackle ‘Disease’ of Fake Degrees sara.custer@in… Fri, 07/10/2026 - 03:00 AM New technologies are making academic fraud ever more common and sophisticated. Byline(s) Seher Asaf for Times Higher Education
Campuses as Places Sara Brady Fri, 07/10/2026 - 03:00 AM New uses for an existing resource. Byline(s) Matt Reed
States Need Better Adult Learner Strategies Joshua.Bay Fri, 07/10/2026 - 03:00 AM More than 43 million Americans have college credits but no credential. A new report from ReUp Education outlines how states can improve efforts to re-engage them. Byline(s) Joshua Bay
Seattle University Dismisses Professor Who Publicly Criticized Provost Emma Whitford Fri, 07/10/2026 - 03:00 AM Carmen Rivera said on Instagram she was “disturbed” by how the provost grabbed a Palestinian flag from the hands of a graduate. Two weeks later, her contract was not renewed. Byline(s) Emma Whitford
Springer Nature Takes Back Retraction of Max Planck’s Papers Sara Weissman Fri, 07/10/2026 - 03:00 AM Byline(s) Sara Weissman
PASSHE Commits to Covering Remaining Tuition for Low-Income Students jessica.blake@… Fri, 07/10/2026 - 03:00 AM Byline(s) Jessica Blake
ED Encouraged Institutions to Limit Graduate Loans. They Don’t Want To. Johanna Alonso Fri, 07/10/2026 - 03:00 AM Colleges are debating whether they should limit lending to students in certain programs that are currently considered “professional” by the Education Department. Byline(s) Johanna Alonso Jessica Blake
Faculty and Students Demand Yale Not ‘Cave’ to Trump kathryn.palmer… Fri, 07/10/2026 - 03:00 AM Reports surfaced late last month that Yale University is negotiating a deal with the Trump administration, which is investigating the university’s admissions practices. Byline(s) Kathryn Palmer
The public Philadelphia institution is reducing its workforce for the second year in a row but has seen promising enrollment gains of late.
From young students’ reading progress to new state laws mandating district policies, what did you learn from our recent stories?
Just 13% of districts studied by a Johns Hopkins center were able to rebound to pre-pandemic levels by the 2024-25 school year.
Article URL: https://thezvi.substack.com/p/childhood-and-education-20-phones Comments URL: https://news.ycombinator.com/item?id=48854496 Points: 2 # Comments: 0
Despite challenges, teachers believe AI can broaden students’ horizons.
arXiv:2606.05194v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are increasingly being deployed to make decisions that require trading off near-term gains against long-term consequences, yet little is known about how they internally represent or resolve these tradeoffs. In this work, we causally localize an underlying subgraph for temporal preference in a distilled LLM (Qwen3-4B-Instruct-2507), identifying mid-to-upper-layer nodes through converging evidence from gradient-based attribution and activation patching. We find that the geometry of time horizon is encoded in the residual stream at the expected localized layers. A behavioral analysis reveals that unintervened LLMs discount the future several times less steeply than humans, yet this preference is unstable across contexts, motivating explicit control rather than implicit reliance on training. Finally, we find suggestive evidence that steering vectors can shift temporal preference. Our work demonstrates ho
arXiv:2605.00155v3 Announce Type: replace-cross Abstract: Reinforcement learning from human feedback (RLHF) is a central post-training tool for aligning large language models, but its training reward is only a learned proxy for true human utility. This creates a decision problem under objective misspecification: the policy is optimized against an estimated reward, while deployment performance is governed by an unobserved population preference. The resulting gap leads to reward over-optimization, where proxy reward keeps improving after true quality deteriorates. We propose distributionally robust regret optimization (DRRO) for RLHF with a Wasserstein ambiguity set over reward laws, using promptwise $\ell_p$ distances between reward vectors as transport costs. Unlike standard distributionally robust optimization, which pessimizes worst-case value, DRRO pessimizes worst-case regret relative to the best policy under the same plausible reward perturbation. We show that the expressive-polic
arXiv:2604.22951v2 Announce Type: replace-cross Abstract: Natural language data follows a power-law distribution, with most knowledge and skills appearing at very low frequency. While a common intuition suggests that reweighting or curating data towards a uniform distribution may help models better learn these long-tail skills, we find a counterintuitive result: across a wide range of compositional reasoning tasks, such as state tracking and multi-step arithmetic, training under power-law distributions consistently outperforms training under uniform distributions. To understand this advantage, we introduce a minimalist skill-composition task and show that learning under a power-law distribution provably requires significantly less training data. Our theoretical analysis reveals that power law sampling induces a beneficial asymmetry that improves the pathological loss landscape, which enables models to first acquire high-frequency skill compositions with low data complexity, which in tu
arXiv:2604.12138v3 Announce Type: replace-cross Abstract: This position paper argues that Retrieval-Augmented Generation (RAG) systems exhibit a factual bias-optimizing for epistemic uncertainty reduction while ignoring the aleatoric uncertainty inherent in opinion-rich content. This misalignment demands a paradigm shift in RAG system design. A survey of 34 major RAG benchmarks reveals that only one addresses opinion synthesis, confirming that the bias is structural and embedded in datasets, retrieval-generation objectives, and evaluation metrics alike. Beyond technical limitations, this bias poses risks to transparent and accountable AI. Namely, echo chamber effects that amplify dominant viewpoints, which can lead to opinion manipulation and under-representation of minority voices. We formalize the problem through the lens of uncertainty quantification, showing that factual queries should minimize posterior entropy while opinion queries must preserve it. We derive a unified objective
arXiv:2602.13376v2 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) are increasingly used in document processing pipelines to convert flowchart images into structured code (e.g., Mermaid). In production, these systems process arbitrary inputs for which no ground-truth code exists, making output quality difficult to assess. We propose a reference-free evaluation framework that monitors flowchart image-to-code generation quality at inference time, using only the input image and the generated output. The framework introduces two automated metrics: $\text{Recall}{\text{OCR}}$, which estimates content coverage by extracting text from the input image via OCR as a proxy reference, and $\text{Precision}{\text{VE}}$, which detects hallucinated elements through Visual Entailment against the original image. Their harmonic mean, $\text{F1}{\text{OCR-VE}}$, provides a unified quality score. Validation on the FlowVQA dataset shows strong agreement with ground-truth metrics (avera
arXiv:2602.04718v2 Announce Type: replace-cross Abstract: A central premise in mechanistic interpretability is that meaningful concepts in language models are represented by linear features in activation space. For such features to support reliable interventions, manipulating one feature should not substantially alter the effects of others. In practice, however, feature entanglement leads to interference such that localized interventions can have unintended downstream effects. Motivated by the \textit{Independent Causal Mechanisms} principle, we propose to constrain internal features to be almost orthogonal. We argue that this promotes modular representations amenable to causal intervention. We formalize this problem by characterizing the gap between an idealized isolated intervention and its realized effect on model outputs in terms of feature interference. We upper-bound the propagation of feature interference in terms of the self-coherence of the feature dictionary, and relate this
arXiv:2601.22448v2 Announce Type: replace-cross Abstract: RLVR has become a standard recipe for training LLMs on reasoning tasks with verifiable outcomes, but when rollout generation dominates the cost, efficiency hinges on which prompts are sampled and when. In practice, prompt pools are often static or only weakly coupled to policy progress, so uniform sampling fails to track the moving capability frontier and wastes rollouts on regions that are already solved or still unreachable. Prior methods improve efficiency via filtering, curricula, adaptive rollout allocation, or teacher guidance, but they often assume a fixed pool, which does not support stable on-policy pool growth, or they introduce additional teacher cost and latency. In this work, we propose HeaPA (Heap Sampling and On-Policy Query Augmentation), which maintains a bounded, evolving pool, tracks the frontier with heap-based boundary sampling, grows the pool via on-policy augmentation under lightweight asynchronous validat
arXiv:2508.11214v2 Announce Type: replace-cross Abstract: Explanations of cognitive behavior often appeal to computations over representations. What does it take for a system to implement a given computation over suitable representational vehicles within that system? We argue that the language of causality -- and specifically the theory of causal abstraction -- provides a fruitful lens on this topic. Drawing on current discussions in deep learning with artificial neural networks, we illustrate how classical themes in the philosophy of computation and cognition resurface in contemporary machine learning. We offer an account of computational implementation grounded in causal abstraction, and examine the role for representation in the resulting picture. We argue that these issues are most profitably explored in connection with generalization and prediction.
arXiv:2508.09767v3 Announce Type: replace-cross Abstract: We propose UtterTune, a lightweight method for adapting a multilingual text-to-speech (TTS) system built on a large language model (LLM). It improves control of pronunciation in the target language while preserving performance in the others. Although LLM architectures have enabled TTS models to achieve remarkable naturalness, accurately modeling grapheme-to-phoneme (G2P) mapping and prosody remains challenging, especially when the model omits an explicit G2P module and directly processes minimally encoded text (e.g., byte-pair encoding). UtterTune leverages low-rank adaptation to enable the control of segmental pronunciation and pitch accent at the phoneme level for Japanese speech, the target language in this paper, while maintaining naturalness and speaker similarity in a zero-shot setting. Objective and subjective evaluations confirm its effectiveness.
arXiv:2607.05583v2 Announce Type: replace Abstract: Contemporary language models are dominated by the transformer architecture, which leverages self-attention mechanisms to enable more efficient, parallelized training across a wide set of documents and corpora. This has allowed transformers to effectively model data across a wide range of modalities and contexts. However, transformers, along with their conventional counterparts such as recurrent neural networks (RNNs) and convolutional neural networks (CNNs), often struggle to maintain efficiency when processing long contexts. We introduce ResonatorLM, a new mechanism that replaces attention with a physics-derived alternative. ResonatorLM treats token sequences as a single, driven one-dimensional latent field and replaces attention dot products with causal functions of damped resonators. We implement ResonatorLM on a traditional network architecture and test it on standard long-context modeling tasks. We find that in a small, 6M matche
arXiv:2606.05622v2 Announce Type: replace Abstract: Planning for real-world problems by language models often involves both world and user constraints, which may not be fully specified upfront and are progressively disclosed through interaction. However, existing benchmarks still underexplore adaptive planning under such progressively revealed dual constraints. To address this gap, we introduce AdaPlanBench, a dynamic interactive benchmark for evaluating whether Large Language Model (LLM) agents can adaptively plan and re-plan under progressively revealed world and user constraints. AdaPlanBench is built on 307 household tasks, with a scalable constraint construction pipeline that augments each task with dual constraints. At runtime, agents interact with the environment in a multi-turn protocol where hidden constraints are revealed only when the agent proposes a plan that violates them, requiring iterative plan revision under accumulating feedback. This makes planning challenging, as a
arXiv:2605.07409v2 Announce Type: replace Abstract: Natural Language Processing is rapidly evolving into a primary instrument for Computational Social Science, with researchers increasingly using embeddings to measure latent constructs such as novelty, creativity, and bias. However, this transition faces a fundamental validity challenge: the ''Proxy Presumption,'' or the reliance on geometric properties (e.g., cosine distance) as direct measures of social concepts. We argue that without explicit validation, unsupervised representations remain entangled mixtures of the target construct ($C$) and confounding attributes ($Z$) like topic, style, and authorship. To bridge the gap between semantic embeddings and valid social measures, we introduce the Construct Validity Protocol (CVP). Drawing on causal representation learning and psychometrics, the CVP offers a rigorous pipeline from conceptualization to quantitative verification. We further propose Counterfactual Neutralization, a novel me
arXiv:2604.13356v3 Announce Type: replace Abstract: Mechanisms for continued self-improvement of language models without external supervision remain an open challenge. We propose Peer-Predictive Self-Training (PST), a label-free fine-tuning framework in which multiple language models improve collaboratively by using a cross-model aggregate response as an internal training signal. Given a prompt, models generate responses sequentially; the final aggregated answer, which is often more reliable than individual responses in practice, serves as an internal reference for learning. We measure how informative each intermediate response is about the aggregate using pointwise mutual information (PMI), and use this signal to scale self-training updates: responses already aligned with the aggregate receive smaller updates, while less informative or misaligned responses receive larger ones. On mathematical reasoning benchmarks, including SimulEq, MATH-500-Numeric, and MultiArith, PST improves exact
arXiv:2604.07102v2 Announce Type: replace Abstract: Activation-based steering enables inference-time personalization of large language models, but its effects in educational applications are not well understood. We study activation-based persona vectors representing seven character traits in short-answer generation and automated scoring on the ASAP-SAS benchmark, across three language models spanning dense and mixture-of-experts architectures. Persona steering lowers answer quality overall, with much larger effects on open-ended English Language Arts (ELA) prompts than on factual science prompts. Interpretive and argumentative tasks are particularly sensitive, showing up to 11$\times$ larger degradation. On the scoring side, we observe predictable valence-aligned calibration shifts: ``evil'' and ``impolite'' scorers grade more harshly, while ``good'' and ``optimistic'' scorers grade more leniently. ELA tasks are 2.5-3$\times$ more susceptible to scorer personalization than science task
arXiv:2602.23440v4 Announce Type: replace Abstract: Reinforcement learning has emerged as an effective paradigm for training large language models to interleave reasoning with search engine calls. However, existing approaches face a fundamental credit assignment problem: methods like Search-R1 assign a single outcome reward to the entire multi-step trajectory, providing no signal about which reasoning or retrieval decisions were responsible for success or failure. Process-reward methods such as StepSearch introduce step-level supervision but still sample complete trajectories independently, so advantage estimates at any given step are contaminated by the randomness of all other steps. We propose SLATE (Step-Level Advantage estimation for Truncated Exploration), which addresses both problems through two complementary ideas. First, truncated step-level sampling generates k continuations from a shared prefix, isolating all variation to a single decision point. We prove this reduces the va
arXiv:2601.22588v2 Announce Type: replace Abstract: Large language models (LLMs) are widely used as reference-free evaluators via prompting, but this "LLM-as-a-Judge" paradigm is costly, opaque, and sensitive to prompt design. In this work, we investigate whether smaller models can serve as efficient evaluators by leveraging internal representations instead of surface generation. We uncover a consistent empirical pattern: small LMs, despite with weak generative ability, encode rich evaluative signals in their hidden states. This motivates us to propose the Semantic Capacity Asymmetry Hypothesis: evaluation requires significantly less semantic capacity than generation and can be grounded in intermediate representations, suggesting that evaluation does not necessarily need to rely on large-scale generative models but can instead leverage latent features from smaller ones. Our findings motivate a paradigm shift from LLM-as-a-Judge to Representation-as-a-Judge, a decoding-free evaluation s
arXiv:2601.10504v2 Announce Type: replace Abstract: As Large Language Models (LLMs) increasingly operate as Deep Research (DR) Agents capable of autonomous investigation and information synthesis, reliable evaluation of their task performance has become a critical bottleneck. Current benchmarks predominantly rely on static datasets, which suffer from several limitations: limited task generality, temporal misalignment, and data contamination. To address these, we introduce DR-Arena, a fully automated evaluation framework that pushes DR agents to their capability limits through dynamic investigation. DR-Arena constructs real-time Information Trees from fresh web trends to ensure the evaluation rubric is synchronized with the live world state, and employs an automated Examiner to generate structured tasks testing two orthogonal capabilities: Deep reasoning and Wide coverage. DR-Arena further adopts Adaptive Evolvement Loop, a state-machine controller that dynamically escalates task comple
arXiv:2509.26076v2 Announce Type: replace Abstract: As the mathematical capabilities of large language models (LLMs) improve, it becomes increasingly important to evaluate their performance on research-level tasks at the frontier of mathematical knowledge. However, existing benchmarks are limited, as they focus solely on final-answer questions or high-school competition problems. To address this gap, we introduce IMProofBench, a private benchmark consisting of 77 peer-reviewed problems developed by expert mathematicians. Each problem requires a detailed proof and is paired with subproblems that have final answers, supporting both an evaluation by human experts and a large-scale quantitative analysis through automated grading. Furthermore, unlike prior benchmarks, the evaluation setup simulates a realistic research environment: models operate in an agentic framework with tools like web search for literature review and mathematical software such as SageMath. Our results show that current
arXiv:2506.15138v2 Announce Type: replace Abstract: Tokenization directly affects the inference efficiency of large language models, since fragmented tokenization increases sequence length and generation cost. Although longer, multi-word tokens can reduce fertility, naively adding them often degrades language model performance. We propose Thunder-Tok, a subword tokenizer that reduces fertility while preserving downstream performance. Thunder-Tok first constructs a large seed vocabulary from corpus substrings and filters structurally incomplete candidates, including invalid Unicode byte fragments and word-boundary violations. It then prunes the seed vocabulary using a likelihood-based token score derived from a uniform Jensen lower bound of the training-data probability. Experiments show that Thunder-Tok reduces fertility by approximately 25% in English and 9% in Korean compared with the standard BPE tokenizer while maintaining competitive performance.
arXiv:2505.23842v5 Announce Type: replace Abstract: Large Language Models (LLMs) increasingly power search engines and AI assistants that retrieve and summarize content from many sources. By serving answers directly, these systems obscure the original content creators' contributions, threatening the compensation that sustains a healthy content ecosystem. We frame this as a problem of fair document valuation and compensation, and propose a framework based on the Shapley value. Because exact Shapley computation is prohibitively expensive at scale, we develop Cluster Shapley, an approximation that groups semantically similar documents via LLM embeddings and computes Shapley values at the cluster level, with formal bounds on both the approximation error and the induced revenue-attribution error. On Amazon product review data, off-the-shelf approximations such as Monte Carlo sampling and Kernel SHAP perform suboptimally in LLM settings, whereas Cluster Shapley substantially improves the eff
arXiv:2502.15543v4 Announce Type: replace Abstract: Large language models (LLMs) integrated with retrieval-augmented generation (RAG) have improved factuality by grounding outputs in external evidence. However, they remain susceptible to unfaithful generation, where outputs contradict retrieved context despite its relevance and accuracy. Existing approaches aiming to improve faithfulness primarily focus on enhancing the utilization of external context, but often overlook the persistent influence of internal parametric knowledge during generation. In this work, we investigate the internal mechanisms behind unfaithful generation and identify a subset of mid-to-deep feed-forward networks (FFNs) that are disproportionately activated in such cases. Building on this insight, we propose Parametric Knowledge Muting through FFN Suppression (ParamMute), a framework that improves contextual faithfulness by suppressing the activation of unfaithfulness-associated FFNs and calibrating the model towa
arXiv:2311.00296v2 Announce Type: replace Abstract: Because most scientific literature data are unlabeled, semantic representation learning based on unsupervised graphs has become crucial. To enrich scientific-literature features, this paper proposes a semantic representation learning method based on adaptive features and graph neural networks. By introducing adaptive feature processing, scientific-literature features are considered globally and locally. The graph attention mechanism weights and aggregates features of scientific documents connected by citation relations, so that correlations among different documents can be expressed more effectively. In addition, an unsupervised graph neural network semantic representation learning method is proposed. By comparing the mutual information between positive and negative local semantic representations of scientific literature and the global graph semantic representation in the latent space, the graph neural network captures local and globa
arXiv:2607.08716v1 Announce Type: cross Abstract: In long-horizon tasks, decision-relevant state is often scattered across an expanding trajectory, while the action agent must surface it and act. As trajectories grow, task requirements, environment facts, prior attempts, diagnoses, and open subgoals can be buried in the context window or pushed beyond it, failing to influence decisions when needed. We call this failure mode "behavioral state decay". We study memory as an active intervention mechanism rather than passive retrieval. A separate memory agent runs alongside an unmodified action agent, updating a structured memory bank from the recent trajectory and deciding whether to inject a memory-grounded reminder or remain silent. The module is plug-and-play with frontier action agents and existing agent harnesses. Across Terminal-Bench 2.0 and $\tau^2$-Bench, it improves pass@1 for both weaker and stronger action agents, with gains of +8.3 pp on Terminal-Bench and +6.8 pp on $\tau^2$-
arXiv:2607.08625v1 Announce Type: cross Abstract: Consumer-facing health chatbots powered by large language models (LLMs) are increasingly used for symptom assessment. However, chatbot development and evaluation often rely on cooperative, articulate, simulated patients. We analysed 2,053 real patient-chatbot conversations and found that communication patterns and expression of emotions vary widely across users. We developed a patient simulator that separately models clinical content, emotional state, conversational strategy, and communication style. In a Turing-inspired evaluation of realism with 15 human graders, simulated conversations were nearly indistinguishable from real ones, with human graders achieving an accuracy of 55%. We used five distinct patient personae, across 1,164 clinician-graded cases, to evaluate the performance of four LLMs in urgency assessment. We found that communication style can significantly alter triage outcomes. Patient-centred conversational artificial i
arXiv:2607.08540v1 Announce Type: cross Abstract: Conversational information retrieval is challenging since it requires the consideration of the conversation history which potentially gives rise to topic shifts and coreference resolution across previous turns. To address these challenges, previous work mainly rely on traditional fine-tuning of ad-hoc retrievers on conversational datasets or extrapolates their generalizability through multi-tasking. However, this mainstream approach is costly - since it requires model re-training - and exhibits catastrophic forgetting, where the model loses its foundational ad-hoc retrieval performance. In this paper, we fill this gap by introducing model merging as a training-free strategy enabling the design of a single retrieval model that operates across both ad-hoc and conversational settings with no additional fine-tuning. We conduct experiments using linear and non-linear parameter-wise merging strategies - namely Model Soup and Slerp - on standa
arXiv:2607.08497v1 Announce Type: cross Abstract: Recent unified multimodal models show a single architecture can jointly perform vision/language understanding and image generation/editing. However, they repeatedly feed all historical visual and textual inputs into a shared context window, limiting long-horizon multimodal dialogue due to visual token explosion and unreliable cross-turn referencing. We propose a Cognitive-structured Multimodal Agent that externalizes visual information into an Episodic Visual Memory and selectively reactivates relevant episodes during reasoning. The agent consists of a Perceptual Abstraction Engine for structured visual abstraction, a Cognitive Retrieval Engine for cross-turn memory retrieval, and a Multimodal Executive Controller for autonomous task inference and action planning. To address the lack of turn-level retrieval supervision in existing datasets, we develop a Unified Scenario Engine that programmatically generates structured multi-turn conver
arXiv:2607.08493v1 Announce Type: cross Abstract: Subjective NLP tasks often exhibit systematic annotator disagreement, requiring models that represent uncertainty rather than collapse it. We introduce Ensemble Diversity Optimization (EDO), a prediction-space framework that jointly optimizes ensemble weights, effective cardinality, and calibration through a unified differentiable objective. EDO learns ensemble composition and size end-to-end via Gumbel-Softmax relaxation and incorporates a signed diversity regularizer, tuned on validation data, to steer optimization toward either preserving or suppressing disagreement. This regularization prevents ensemble collapse and enables controlled navigation of the utility-calibration trade-off. The framework integrates a soft F1 surrogate, class-weighted cross-entropy to address imbalance, and reliability-weighted diversity to regulate intra-ensemble variability. Experiments on four subjective text-classification benchmarks (ArMIS, ConvAbuse, H
arXiv:2607.08395v1 Announce Type: cross Abstract: Persistent AI agents extend large language models (LLMs) beyond single-turn interaction into long-lived software systems. Unlike traditional chat assistants, unsafe content in these agents can propagate through persistent state, reusable skills, and tool-mediated interactions, creating a substantially larger semantic attack surface. We observe that most security-critical interactions in such agents are transmitted through natural-language token flows, including memory updates, tool arguments, retrieved files, and inter-component communications. This observation enables a new security formulation: unsafe behavior can be intercepted as risky semantic flows before reaching privileged runtime sinks. Based on this insight, we propose TokenWall, a runtime defense framework that acts as a semantic firewall over agent token flows. TokenWall performs boundary-aware semantic auditing over these flows, constructing structured source-sink audit rec
arXiv:2607.08393v1 Announce Type: cross Abstract: Fine-tuning LLMs to inject new knowledge faces a critical challenge: LLMs can quickly memorize new facts, yet fail to use them for downstream reasoning tasks. We formalize this failure as the \textit{\textbf{Knowing--Using Gap}}, characterized by an accuracy gap and a temporal lag between memorization and generalization. To understand this phenomenon, we fine-tune LLMs with unseen knowledge and monitor the spatial permeation dynamics of the knowledge internally using a novel intervention technique called self-patching. Self-patching identifies activation locations where relocating representations substantially improves failed generalization cases. These results are consistent with a knowledge-circuit misalignment hypothesis: memorized representations can exist internally but may not be routed to computation-effective layers. To demonstrate the practicality of this diagnostic finding, we design a simple heuristic strategy which recovers
arXiv:2607.08268v1 Announce Type: cross Abstract: High-volume structured extraction pays a large model's latency on every item, so distilling the task into a small on-device model is attractive: comparable output at a fraction of the time and cost. We measure what that distillation actually delivers, per sub-task. Each news article is mapped to one JSON object with a short summary and five categorical labels. We distill an 8B reasoning teacher (deepseek-r1:8b) into a 0.6B student (Qwen3-0.6B; QLoRA, three seeds), and add two teacher controls: a same-size non-reasoning teacher and a larger managed pipeline. A blinded, reference-free, three-judge panel scores every arm against the full article, alongside two non-distillation baselines, few-shot prompting and constrained decoding. The student runs at about 0.8 s per article against the teacher's 39 s, and recovers 58% of the base-to-teacher gap on summary quality, beating its primary baseline (constrained decoding) by +16.8 points and few