The paper investigates the limits of self-evolution in LLM reasoning under closed-loop settings, finding that while self-improvement is significant, it consistently falls short of perfect oracle super…

The paper introduces Feedback Distillation, a novel training method that uses a language model's privileged feedback to provide token-level supervision, significantly improving complex reasoning tasks…

This paper proposes a post-training framework called Retrieval-Augmented Reinforcement Fine-Tuning (RA-RFT) to teach language models to reason by analogy.

This paper synthesizes over 150 scattered studies and reports to provide the first comprehensive primer on post-training reasoning data, organizing the field around data objects, utility, construction…

The paper introduces Reasoning Exposure Prompting (REP), a method that demonstrates that even when LLMs hide their internal reasoning steps from users, useful reasoning supervision can still be elicit…

The paper introduces Reasoning Exposure Prompting (REP), a method that demonstrates that even when LLMs hide internal reasoning traces from users, useful reasoning supervision can still be elicited th…

ThinkSwitch introduces a low-compute co-training procedure that distills the reasoning benefit of large language models into weights, significantly improving performance on specific reasoning tasks.

DenoiseRL is a novel reinforcement learning framework that improves reasoning in large language models by optimizing directly from the failures and incorrect reasoning traces of weak models, eliminati…

The paper introduces HRBench, a unified and comprehensive evaluation framework for systematically benchmarking and comparing various thinking-mode switching strategies in hybrid-reasoning LLMs.

The paper proposes Skill-Conditioned Gated Self-Distillation (SGSD), a novel framework that uses retrieved, potentially noisy skills to guide LLM reasoning, achieving state-of-the-art performance on m…

The paper introduces TRACE, a novel metric that evaluates the logical structure of LLM reasoning (CoT) by integrating Toulmin's argumentation theory, demonstrating that sound reasoning structure corre…

The paper proposes a novel, efficient method for checking the factuality of claims generated by LLMs by framing it as a true/false reading comprehension task and incorporating explicit test-taking str…

DenseSteer is a training-free inference-time framework that improves the math reasoning capabilities of small language models by steering their internal representations toward a 'Dense Reasoning' patt…

LongTraceRL addresses long-context reasoning challenges by generating highly challenging training data and introducing a fine-grained rubric reward, significantly improving evidence-grounded reasoning…

The paper introduces RefMem-Bench, a new benchmark for measuring reflective memory in long-horizon dialogue, and proposes REMIND, a framework that significantly improves models' ability to synthesize…

The paper proposes Visual Gradient Steering (VGS), a method that decomposes the distillation loss into language and visual components and steers the optimization to prioritize visual grounding, signif…

The paper compares verbalized feature attributions and self-generated rationales for explaining model behavior, finding that the format and granularity of the explanation significantly affect its abil…

The paper introduces Reasoning in Memory (RiM), a latent reasoning method that replaces autoregressive token generation with fixed memory blocks to enable compute-efficient internal working memory for…

The paper proposes an unsupervised Reinforcement Learning approach that enforces cross-lingual self-consistency to significantly enhance the multilingual reasoning capabilities of large language model…