Causal Evidence that Language Models use Confidence to Drive Behavior

TL;DR

This paper provides causal evidence that language models utilize internal confidence signals, including verbal cues, to control their behavior, such as deciding when to abstain, demonstrating structured metacognitive control.

Contribution

It introduces a four-phase paradigm to causally demonstrate how LLMs use confidence signals, including verbal cues, for abstention decisions, advancing understanding of model metacognition.

Findings

Models apply an implicit confidence threshold for abstention.

Boosting confidence signals decreases abstention rates.

Verbal confidence predicts abstention independently of output distribution.

Abstract

Metacognition -- assessing the quality of one's own cognitive performance -- guides adaptive behavior across species. Substantial research demonstrates that confidence signals can be extracted from language model outputs, yet a fundamental question remains: do models actually use these signals to control behavior, such as deciding whether to answer or abstain? To investigate, we developed a four-phase paradigm. Phase~1 elicited baseline confidence estimates without an abstention option. Phase~2 revealed that LLMs apply an implicit threshold to internal confidence when deciding to abstain, with confidence effect sizes approximately an order of magnitude larger than alternative mechanisms. Phase~3 provided direct causal evidence through activation steering: boosting or suppressing confidence signals correspondingly decreased or increased abstention rates. Phase~4 extended this by systematically varying instructed thresholds, demonstrating that LLMs actively deploy confidence signals to implement abstention policies. Critically, beyond calibrated log-probability based confidence derived from the output distribution, verbal confidence independently predicted abstention across all models, despite being objectively less discriminatory of answer correctness. Activation decoding at the last pre-answer token further showed that both observable measures are lossy readouts of a richer internal representation. Together, these results suggest that abstention is not fully captured by the strength of evidence in the output distribution alone, but is better explained by the joint operation of a multidimensional internal confidence representation and threshold-based policies -- consistent with structured metacognitive control in LLMs, a capacity of growing importance as models transition to autonomous agents that must recognize their own uncertainty.