Analyzing and Improving Chain-of-Thought Monitorability Through Information Theory

Usman Anwar, Tim Bakker, Dana Kianfar, Cristina Pinneri, Christos Louizos

Keywords: chain-of-thought monitoring, mutual information, AI Safety, reasoning, LLMs

TL;DR: CoT monitorability can be improved via training with specific objectives

Abstract:

Chain-of-thought (CoT) monitors are LLM-based systems that analyze reasoning traces to detect when outputs may exhibit attributes of interest, such as test-hacking behavior during code generation.
In this paper, we use information-theoretic analysis to show that although a task requiring CoT satisfies the necessary condition for CoT monitorability, it does not constitute a sufficient condition.
We identify two sources of approximation error that may undermine the performance of CoT monitors in practice: _information gap_ which intuitively measures the extent to which the monitor can extract the information available in CoT, and _elicitation error_ which measures the extent to which the monitor approximates the optimal monitoring function.
We further demonstrate that CoT monitorability can be systematically improved through targeted training objectives.
To this end, we propose two complementary approaches: (a) an oracle-based method that directly rewards the monitored model for producing CoTs that maximize monitor accuracy, and (b) a more practical, label-free approach that maximizes conditional mutual information between outputs and CoTs.
Across multiple different environments, we show both methods significantly improve monitor accuracy while preventing CoT degeneration even when training against a monitor, thereby mitigating reward hacking when the task reward is imperfectly specified.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 13673