Frontier LLMs struggle to discriminate data uncertainty from model uncertainty even when accurate, but a new benchmark and lightweight RL strategy improve attribution without sacrificing answer accuracy.
Octothinker: Mid-training incentivizes reinforcement learning scaling.arXiv preprint arXiv:2506.20512, 2025b
4 Pith papers cite this work, alongside 10 external citations. Polarity classification is still indexing.
4
Pith papers citing it
10
external citations · OpenAlex
fields
cs.CL 4verdicts
UNVERDICTED 4representative citing papers
LMs systematically inflate expressed certainty during rewriting, affecting up to 75% of outputs with a 1.5-2x bias toward increasing rather than decreasing certainty, and the effect compounds over iterations.
BAG prompts LLMs to reason over K sampled responses for strategy selection in multi-turn ambiguous QA, improving accuracy and faithfulness to uncertainty over baselines across six models.
MASH uses RL with a pay-per-search reward to make LLMs seek external help only when needed, improving multi-hop QA accuracy by 7.6% and enabling competitive abstention without pre-defined knowledge boundaries.
citing papers explorer
-
Beyond "I Don't Know": Evaluating LLM Self-Awareness in Discriminating Data and Model Uncertainty
Frontier LLMs struggle to discriminate data uncertainty from model uncertainty even when accurate, but a new benchmark and lightweight RL strategy improve attribution without sacrificing answer accuracy.
-
From `May' to `Is': Certainty Distortion in Language Model Rewriting
LMs systematically inflate expressed certainty during rewriting, affecting up to 75% of outputs with a 1.5-2x bias toward increasing rather than decreasing certainty, and the effect compounds over iterations.
-
Clarify, Abstain or Answer? Strategising in Conversation with Belief-Augmented Generation
BAG prompts LLMs to reason over K sampled responses for strategy selection in multi-turn ambiguous QA, improving accuracy and faithfulness to uncertainty over baselines across six models.
-
MASH: Modeling Abstention via Selective Help-Seeking
MASH uses RL with a pay-per-search reward to make LLMs seek external help only when needed, improving multi-hop QA accuracy by 7.6% and enabling competitive abstention without pre-defined knowledge boundaries.