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.
Signal 37 demonstrates a working method (SMRC) for aligning LLM reasoning to student error patterns, and signal 53 shows on-policy self-distillation can refine such models cheaply at inference time. Meanwhile signal 16 confirms that standard LLMs still hallucinate in high-stakes educational contexts, creating a clear gap that a specialized, student-aligned correction system can fill.
Current AI math tutors detect errors but correct them using model-centric reasoning patterns that don't match how students actually make mistakes, producing explanations that students cannot follow or learn from.
K-12 and undergraduate math students, teachers seeking AI tutoring tools, and edtech platforms building homework help products
A math tutoring system that fine-tunes LLMs specifically to mirror student reasoning patterns when diagnosing and correcting errors—not just flagging the wrong answer but tracing the student's own logical path and intervening at the precise conceptual misstep. The system uses student-aligned correction models (per signal 37's SMRC approach) combined with on-policy self-distillation (signal 53) so the tutor continuously improves its error-correction accuracy against real student solution traces rather than model-generated ones.
The real-world evidence the pipeline drew on to generate this idea.
arXiv:2511.14684v2 Announce Type: replace Abstract: Large language models (LLMs) often make reasoning errors when solving mathematical problems, and how to automatically detect and correct these errors has become an important research direction. However, existing approaches \textit{mainly focus on self-correction within the model}, which falls short of the "teacher-style" correction required in educational settings, \textit{i.e.}, systematically guiding and revising a student' s problem-solving process. To address this gap, we propose \texttt{SMRC} (\textit{\underline{S}tudent \underline{M}athematical \underline{R}easoning \underline{C}orrection}), a novel method that aligns LLMs with student reasoning. Specifically, \texttt{SMRC} formulates student reasoning as a multi-step sequential decision problem and introduces Monte Carlo Tree Search (MCTS) to explore optimal correction paths. To reduce the cost of the annotating process-level rewards, we leverage breadth-first search (BFS) guid
arXiv:2608.28306v1 Announce Type: cross Abstract: On-policy self-distillation (OPSD) improves reasoning by training a problem-only student on its own rollouts using dense token-level supervision from a privileged teacher that also sees a reference solution. However, standard OPSD treats the teacher distribution as a fixed target along the student's rollout and updates only the student %, although -- even though privileged conditioning does not guarantee that the teacher always provides the most appropriate target for problem-only reasoning. This one-way supervision can therefore misdirect the student when the teacher distribution is misaligned with valid student reasoning. We therefore introduce Verifier-Informed Student-to-Teacher Adaptation (VISTA), which preserves the standard OPSD student update while using outcome-verified rollouts to adapt the teacher toward the student distribution. Within each verified rollout, VISTA further restricts this adaptation to the top-$k$ positions wi
arXiv:2606.21359v2 Announce Type: replace Abstract: Large language models (LLMs) are increasingly used to communicate and explain scientific concepts, yet their tendency to hallucinate poses significant risks in this high stakes use-case. Prior scientific hallucination evaluation work remains largely restricted to the biomedical domain, treats hallucination as a binary task, and has not examined the growing family of scientifically fine-tuned LLMs. We address these gaps with SciFactCheck, a benchmark of 2,500 prompts across five scientific domains, paired with a modular evaluation framework targeting three factuality hallucination types: unverifiability, overclaim, and attribution. Using a controlled minimal-pairing design, we evaluate 18 LLMs by comparing each scientifically fine-tuned model against its general-purpose base. Our results indicate that 1. Scientifically fine-tuned models exhibit degraded factual reliability across all hallucination types and scientific domains, and 2. F
arXiv:2605.30504v2 Announce Type: replace Abstract: LLM benchmark labels are frozen at release and silently propagated into downstream benchmarks, errors and all. We introduce an Item Response Theory-based indicator that surfaces likely mislabels at 95% precision in the top 200 examples across seven preference and multiple-choice benchmarks using responses from 114 models, outperforming a supervised classifier. We trace these errors to mechanical labeling heuristics, upstream annotation mistakes inherited unchanged from source datasets, and fundamentally ambiguous items without a defensible single label. The same model fit reveals that reward models specialize in stylistic preference rather than factual knowledge, and identifies one frontier reward model that agrees with detected mislabels at 78% accuracy versus 38% for its peers, consistent with benchmark contamination or benchmark-specific over-optimization.
arXiv:2511.08577v4 Announce Type: replace Abstract: Improving the reasoning abilities of Large Language Models (LLMs), especially under parameter constraints, is crucial for real-world applications. Looped transformers address this by performing multiple latent iterations to refine each token beyond a single forward pass. However, we identify a latent overthinking phenomenon: most token predictions are already correct after the first pass, but are sometimes revised into errors in later iterations. We ask whether selectively skipping latent iterations can improve accuracy, and reveal significant potential with an oracle iteration policy that boosts performance by up to 7.3%. Motivated by this, we propose Think-at-Hard (TaH), a looped transformer optimized for selective iteration. TaH employs a lightweight neural decider to trigger latent iteration, only at tokens likely to be incorrect after the standard forward pass. During latent iterations, depth-aware Low-Rank Adaptation (LoRA) modu