arXiv:2607. 00276v1 Announce Type: cross Abstract: Current large-language-model (LLM) physics benchmarks are usually scored by answer accuracy, which cannot distinguish genuine reasoning from recall of familiar problem patterns and reveals little about where a model's reasoning breaks down.
arXiv:2608. 16795v1 Announce Type: cross Abstract: Systems that generate scientific research questions are evaluated today by expert scores, LLM-as-judge ratings, or curated case studies -- all subjective, none falsifiable.
arXiv:2606. 27383v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as research assistants, yet it remains unclear whether they can calibrate research takeaways to the strength and scope of the supporting evidence.
arXiv:2605.26087v2 Announce Type: replace-cross Abstract: Frontier LLMs now perform strongly across a wide range of physics evaluations, but it is hard to disentangle genuine reasoning from recall of...
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).
arXiv:2610.01471v1 Announce Type: cross Abstract: Large language models now generate code, documentation, and analyses, and are increasingly used to review such output. We ask when a second review by...
arXiv:2608. 11415v1 Announce Type: cross Abstract: Large language models are being proposed as agents in scientific workflows, in domains where no downstream verifier exists.
arXiv:2606. 09046v1 Announce Type: new Abstract: Useful audits reveal not only how often a model fails, but also where its failures concentrate.
EviScope is a new paired counterfactual benchmark that evaluates grounded language models by fixing the question while manipulating evidence—adding, removing, distracting, or contradicting it. The v1.1 dataset includes 40 four‑condition quartets with repaired counterfactual claims and span‑level support labels for automated assessment. Experiments on Qwen2.5‑7B, Llama 3.1 8B, and Gemini 3.5 Flash show that paired metrics reveal grounding behaviors hidden by simple answer accuracy, such as unsupported answers, conflict blindness, and incorrect non‑answer actions.
The paper proposes a distributional theory explaining how large language models (LLMs) incorporate external evidence into their decision-making process. It identifies three key predictions: (1) evidence is more persuasive when it aligns with the model’s prior beliefs, (2) models more readily accept errors from their own internal processes than from external sources, and (3) the same evidence can improve weaker models while harming stronger ones. Extensive experiments across ten million trials, twelve LLMs from four families, and eight domains—including quantum mechanics, physics, genetics, and molecular biology—confirm these predictions and reveal that evidence integration occurs late in the network as a structured sequence of steps rather than through a simple trust metric.
The paper introduces inexpensive, scalable methods for evaluating language model behavior across different vendors and releases. By running identical public stimuli on a cross‑vendor panel and analyzing transcripts via exact match, LLM‑coded codebooks, or instrumented environments, the authors can quantify model responses at a cost of a few dollars per model. Applying these tools to four years of releases reveals patterns of convergence, resistance, positional stability, and compliance that vary by generation, lab, and harness.
The paper investigates how multi‑modal world models can produce inconsistent outputs across different modalities, such as a video showing a ball not rebounding while a text description indicates it should. It defines two types of misalignment—internal (between modalities) and external (against a physical environment)—and introduces a physics‑grounded pipeline to measure these discrepancies. Experiments across multiple settings reveal that while the model’s language output matches the true environment, its video output frequently disagrees, indicating current unified backbones struggle with simultaneous reasoning, consistency, and physical fidelity.
The paper investigates whether large language models’ reasoning traces truly contain early, informative signals or merely reflect budget and difficulty confounds. Using a restart‑controlled truncation probe, the authors compare continuation success rates against from‑scratch restart curves across 178 problem‑model pairs, finding that only one case shows prefix‑limited success and that continuing a model’s own prefix generally outperforms restarting. A difficulty‑controlled test and two generation‑free analyses reveal that early internal signals do not carry outcome information beyond a problem‑difficulty baseline, underscoring the need for proper counterfactual controls.