The study investigates whether the reasoning steps a language model writes are causally responsible for its answers. Using a causal intervention method on the activation stream, the authors find that for Qwen3-4B, about 77% of stated steps are causally load‑bearing, while behavioral tests overestimate this by roughly 11 percentage points. The faithfulness of reasoning decreases with model size and depth of reasoning, especially for the smaller Qwen3-1.7B.
By Abhiram Bhupatiraju, Rayan Nyaupane
arXiv:2607. 08065v1 Announce Type: new Abstract: LLM-as-judge (Zheng et al.
By Kaihua Ding
arXiv:2606. 08275v1 Announce Type: cross Abstract: When an LLM agent fails -- issues a refund it should not have, calls the wrong tool, leaks data -- existing tooling answers what happened (observability) or whether it passed (evaluation), but not which step caused the failure.
By Jaineet Shah
The paper investigates how different scoring rules affect the ranking of large language models (LLMs) on the IndiaFinBench benchmark, which contains 406 expert‑annotated items covering regulatory interpretation, numerical reasoning, contradiction detection, and temporal reasoning. Using twelve LLMs evaluated under strict string‑matching, a cross‑model judge, and a composite regime, the authors find that model rankings vary dramatically across regimes, with no significant rank correlation (Spearman ρ = -0.273). Item‑level analysis reveals that a substantial portion of items are answered identically by all models, while a higher‑discrimination band isolates more challenging cases.
By Rajveer Singh Pall
The study investigates how deictic ambiguity—specifically the shifting reference of expressions like "previous"—affects Draft‑Verify‑Revise pipelines that use multiple large language models (LLMs). Using a synthetic dataset of 10 base examples and 21 reasoning‑effort configurations, six LLMs were evaluated for their ability to correctly resolve the ambiguous expression across the draft, verify, and revise stages. Results show wide variance in balanced accuracy, with GPT‑5.2 improving from 0.156 to 0.942 with increased reasoning effort, while Gemini 3 Pro consistently achieved high accuracy above 0.94 even at low reasoning effort, and meta‑evaluators often relied on surface cues when making errors.
By Obinna I. Ekekezie
arXiv:2608. 08514v1 Announce Type: new Abstract: We independently reproduce two recent methods for making large language model (LLM) reasoning more reliable, and stress-test them across domains and models (RPC across four new task domains with Qwen3-8B, LCF across four 7-8B models).
By Minhan Cho, Jimin Kweon
The paper investigates how language‑model judges can make version‑dependent errors when evaluating upgraded agents. Using 35 public coding‑agent submissions, two customer‑service agents, and over a thousand expert‑labeled trajectories, the authors show that fixed judges often reject task‑conditioned error invariance and can incorrectly approve failed patches, especially as agent capability increases. Paired audits of current outputs reduce interval width only marginally, and the study concludes that independent human patch review is still necessary.
By Jiapeng Li
The paper introduces a counterfactual evidence audit to assess how large language model agents respond to different sets of documents. By exposing agents to two mirrored sets of five documents and measuring changes in six downstream decisions, the authors predict how the agents will behave when presented with a full 45‑document context. Across multiple open‑weight model families and tasks, the five‑document audit accurately forecasts full‑context effects, reducing prediction error by 62% and correctly identifying the direction of most material effects.
By Rana Muhammad Usman
The paper compares a non‑generative verifier (Jev) with a frontier LLM (Astra) on 495 human‑labelled claims from LLM‑AggreFact. Both models achieve statistically indistinguishable accuracy, yet Jev is 187‑fold cheaper and admits more unsupported claims. A confidence‑gated cascade that escalates uncertain claims improves cost‑efficiency, though the benefit comes mainly from the escalation share rather than the probability estimates.
By Alden Do Rosario, Kirollos Atef
arXiv:2610.02267v1 Announce Type: new
Abstract: Agent harnesses make many small, typed decisions per task: which model to call, which tool to use, whether retrieved text is relevant, whether an input...
By Jiawei Li
arXiv:2608. 19760v1 Announce Type: cross Abstract: Audited against causal ground truth from executed replay in a single-agent tool environment (ALFWorld), none of the step-level credit signals used to train LLM agents -- LLM-judge scores, outcome-conditioned logprob ratios, or the policy's own confidence -- identifies which steps causally matter better than chance.
By Haiyue Zhang
arXiv:2606. 23767v1 Announce Type: new Abstract: Headline accuracies on the Tuebingen cause-effect pairs are routinely compared across papers even though each is measured under its authors' own protocol -- different pair subsets, weightings, model-selection, and decision rates.
By Wietse Stienstra