Dynamic Rollout Editing reduces overthinking in RL-trained LLMs by editing post-answer continuations in successful rollouts and preferring the edited versions within GRPO groups.
Answer convergence as a signal for early stopping in reasoning.arXiv preprint arXiv:2506.02536
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
verdicts
UNVERDICTED 7representative citing papers
RecurGuard monitors recurrence rate, volume growth, and query progress in exposed reasoning traces to terminate generation on token-consumption attacks, reporting 99% detection on OverThink and 92% on ExtendAttack with near-zero false positives.
RL-trained lightweight controller using answer statistics improves trade-offs among correctness, latency, and total samples in adaptive sampling for LLM test-time scaling.
Conformal risk control with upper and lower thresholds lets LLMs adaptively stop reasoning while guaranteeing a maximum error rate and minimizing token use.
Entropy After </Think> (EAT) enables early exiting in reasoning LLMs by tracking entropy stabilization after a </think> token, cutting token use 12-22% on MATH500 and AIME2025 with no accuracy loss.
TRACE aggregates answer consistency and confidence trajectory over multiple reasoning steps to decide when to halt inference, reducing token usage by 25-30% while keeping accuracy within 1-2% of full reasoning.
DTSR enables large reasoning models to dynamically assess chain-of-thought sufficiency via reflection signals and a sufficiency check, reducing reasoning length by 28.9-34.9% with minimal performance loss on Qwen3 models.
citing papers explorer
-
Entropy After </Think> for reasoning model early exiting
Entropy After </Think> (EAT) enables early exiting in reasoning LLMs by tracking entropy stabilization after a </think> token, cutting token use 12-22% on MATH500 and AIME2025 with no accuracy loss.