arXiv AI

Chain-of-Thought Faithfulness of Reasoning Models Varies with Where and How Preference Cues Are Delivered

arXiv Computation and Language
Oct 8

The Long Road to the Same Answer: Cognitive Bias Under Escalating Reasoning Budgets in Large Language Models

The study investigates whether large language models (LLMs) that allocate extra computation during inference—termed reasoning models—reduce classic decision biases compared to their non‑reasoning counterparts. Using 30 vignettes covering six cognitive biases and varying token budgets up to 8,192 tokens, the authors find that reasoning models are not less biased, and increased deliberation does not reliably diminish bias magnitude. Only anchoring showed a bias in the human direction, while other biases either remained unchanged or moved further from human patterns, suggesting that test‑time reasoning does not guarantee rationality.

By Obada Kraishan
arXiv Computation and Language
Sep 1

Detecting Hidden Chain-of-Thought in Large Language Models with Linguistic, Behavioral, and Mechanistic Indicators

arXiv:2608.29956v1 Announce Type: new Abstract: Large language models often answer complex reasoning questions without revealing intermediate steps, raising whether they reason latently or complete p...

By Armaan Singh, Ryan Trinh Le, Jasmine Kaur, Abdullah Sultan, Edward Lue Chee Lip, Kiran Nijjer, Adnan Ahmed, Vasu Sharma
arXiv Computation and Language
Nov 5

ProcessPrefBench: Evaluating Cross-Task Transfer of Individual Process Preferences from Naturalistic Human-AI Interaction

ProcessPrefBench evaluates how well AI assistants can transfer individual process preferences—such as standards for evidence, depth, and delivery—across different tasks using naturalistic human-AI dialogue data. The benchmark uses 66 source‑target pairs from three months of conversations with two knowledge workers, focusing on applying preferences rather than retrieving them. Experiments show that prompting with full dialogue history or distilled preference instructions improves preference realization by about 0.2 points on a 0‑2 scale, but also increases fabricated‑evidence failures.

By Yiwen Yang, Guanming Liu
arXiv AI
Oct 5

Efficient Reasoning Training Does Not Always Harm CoT Faithfulness and Monitorability

The paper investigates how training large language models to use fewer tokens in chain-of-thought (CoT) reasoning impacts the faithfulness and monitorability of the generated explanations. Three efficiency methods—fixed generation budget, per-example length target, and group-relative length reward—were applied during fine‑tuning, and the resulting models were evaluated on how well their CoT reflects decision processes and whether it signals changes due to input interventions. Results show that while faithfulness generally decreases because models become less consistent, monitorability remains relatively robust, with models still indicating the influence of input changes even when CoT is shortened.

By Samuel Lewis-Lim, Xingwei Tan, Mario Sanger, Zhixue Zhao, Nikolaos Aletras
arXiv Machine Learning
Oct 23

Interpreting Model Reasoning through Behavioral Counterfactuals

The paper introduces a method for analyzing how interventions in a prompt affect a language model’s reasoning by using behavioral counterfactuals. It trains a module to edit the residual stream so that the next-token distributions for a fixed response match between a source and a counterfactual prompt, outperforming standard activation reconstruction. The resulting attribution maps and visualizations help identify causally important locations for counterfactual recovery and guide hypothesis testing with activation patching, as demonstrated in a case study on Gemma-4-E4B-it and DeepSeek-R1-0528-Qwen3-8B.

By Shawn Im, James Oldfield, Adel Bibi, Sharon Li
arXiv Machine Learning
Aug 14

A Probe Direction Is a Property of Its Prompt

arXiv:2608. 13329v1 Announce Type: new Abstract: A model that behaves differently when it senses it is being tested would undermine the evaluations we rely on, so recent work has sought to read that sense directly from a model's activations.

By Valentin No\"el
arXiv AI
Sep 17

Bypassing the Rationale: Causal Auditing of Implicit Reasoning in Language Models

The paper introduces a causal, layerwise audit method called the CoT Mediation Index (CMI) to evaluate whether chain-of-thought (CoT) prompting truly influences a language model’s internal computation. By comparing performance degradation from patching CoT-token hidden states against matched control patches, the authors find that CoT influence is often confined to narrow reasoning windows and can be nearly absent even when the model produces fluent rationales. The study shows that models explicitly tuned for reasoning exhibit stronger mediation, while Mixture-of-Experts models display more distributed mediation, indicating that CoT faithfulness varies across models and tasks.

By Anish Sathyanarayanan, Aditya Nagarsekar, Aarush Rathore