This work uses mechanistic interventions in order to disentangle whether reasoning chains should be viewed as a cause of vs. a justification for the model’s final answer, and finds that the answer to this question is complicated by the fact that the interpretability tools on which the model uses influence the mechanism the model uses, and thus the fidelity of the chain itself.
Results show that capability failures can manifest as distributed, task-dependent changes in the structure of visible reasoning, and that CoT dynamics agnostic to whether the verbalized trace reflects the model's internal computations can help diagnose and correct failures.
Experiments show that CASE achieves a 37\% average per-setting relative improvement in overall CoT faithfulness over the strongest baselines, exhibits stronger cross-dataset faithfulness transfer, and maintains competitive average accuracy.
This synthesis shows that the open frontier of reasoning in driving agents lies in intermediate representations that can be grounded in the real world, coupled to real-time action, and verified under safety-critical systems.
Zheng-Xu Tang, Xiao-Zhou Zhang, Guo-Feng Cui et al.· 0 citations
This work investigates how instruction-tuned Transformer models (LLaMA and Mistral) encode discourse relations in English, with a particular focus on the contrasting relations of causation and antithesis, suggesting asymmetric representation of discourse-based reasoning.
A. Bhattacharyya, Shira Wein· SIGDIAL Conferences· 0 citations
The history and current state of interpretability taxonomized according to the types of causal units utilized, as well as methods used to search over mediators are described.
Aaron Mueller, Jannik Brinkmann, Millicent Li et al.· 0 citations
This work operationalizes the importance of a reasoning step as its advantage: the change in expected reward, e.g., producing the correct final answer, from including that step, estimated via Monte Carlo rollouts, and evaluates whether LLM judges can identify high-advantage steps and finds that sufficiently capable LLM...
Kevin Du, A. Hoyle, L. Ruis et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
