Reasoning Trace
The externalized text of the internal chain-of-thought process that a reasoning model generates before returning its final answer. It is usually returned in a separate field or with special delimiters, but not all vendors expose it to callers.
Timeline (last 90 days)
A reasoning trace is the sequence of intermediate thinking tokens that a large language model outputs step by step before generating a final answer; it improves accuracy on complex tasks but consumes a large number of additional tokens.
Capturing the agent's reasoning trace can serve as direct evidence of security policy effectiveness; if the model explicitly states that its reason for refusal is security-related, it indicates the policy is working.
Models trained on corrupted traces consistently outperformed control groups trained on correct traces in cross-domain tasks, while the reasoning traces they generated were consistently invalid.
The study found that the length of generated reasoning traces is largely independent of the difficulty of the underlying problem.
The study observed a clear lack of correlation between the correctness of answers and the validity of reasoning traces, with models frequently arriving at correct answers even when the reasoning traces are completely invalid.
The article advises developers not to over-trust the readability of reasoning traces, and to focus on the correctness of the final answer rather than the aesthetic quality of the reasoning process.
The researchers argue that assuming reasoning traces possess human-like or algorithmically interpretable semantics is not only unnecessary but potentially misleading.
Reinforcement learning in the post-training stage can improve answer accuracy on both in-distribution and out-of-distribution tasks, but does not consistently improve the validity of reasoning traces.
All Facts (8)
A reasoning trace is the sequence of intermediate thinking tokens that a large language model outputs step by step before generating a final answer; it improves accuracy on complex tasks but consumes a large number of additional tokens.
50%UnverifiedCapturing the agent's reasoning trace can serve as direct evidence of security policy effectiveness; if the model explicitly states that its reason for refusal is security-related, it indicates the policy is working.
50%UnverifiedModels trained on corrupted traces consistently outperformed control groups trained on correct traces in cross-domain tasks, while the reasoning traces they generated were consistently invalid.
50%UnverifiedThe study found that the length of generated reasoning traces is largely independent of the difficulty of the underlying problem.
50%UnverifiedThe study observed a clear lack of correlation between the correctness of answers and the validity of reasoning traces, with models frequently arriving at correct answers even when the reasoning traces are completely invalid.
50%UnverifiedThe article advises developers not to over-trust the readability of reasoning traces, and to focus on the correctness of the final answer rather than the aesthetic quality of the reasoning process.
50%UnverifiedThe researchers argue that assuming reasoning traces possess human-like or algorithmically interpretable semantics is not only unnecessary but potentially misleading.
50%UnverifiedReinforcement learning in the post-training stage can improve answer accuracy on both in-distribution and out-of-distribution tasks, but does not consistently improve the validity of reasoning traces.
50%