Posts
All the articles I've posted.
-
Can Large Reasoning Models Self-Train?
Published: at 11:43 AM
86.73 🤔本文提出Self-Rewarded Training (SRT) 方法，通过模型自一致性驱动强化学习实现无监督数学推理能力提升，初期性能媲美有监督方法，但因奖励黑客问题导致长期训练性能崩溃，并探索了提前停止和课程学习等缓解策略。
-
Hybrid Latent Reasoning via Reinforcement Learning
Published: at 11:43 AM 86.71 🤔本文提出HRPO，一种基于强化学习的混合潜在推理框架，通过门控机制结合离散token和连续隐状态，显著提升了大型语言模型在知识和推理任务上的性能，同时减少了对链式思维数据的依赖。
-
Walk Before You Run! Concise LLM Reasoning via Reinforcement Learning
Published: at 11:13 AM 86.69 🤔本文提出 ConciseR，一种两阶段强化学习框架，通过 GRPO++ 提升推理能力并通过 L-GRPO 优化响应长度，在保持准确性的同时显著减少 CoT 响应长度，优于多个基准数据集上的现有方法。
-
On the Generalization vs Fidelity Paradox in Knowledge Distillation
Published: at 11:23 AM 86.63 🤔本文通过大规模实证分析揭示知识蒸馏（KD）显著提升小型语言模型的零样本推理性能（高达10%），但对大型模型收益有限，且性能提升与推理保真度存在脱节，强调任务专长和适度参数调整的重要性。
-
MiMo: Unlocking the Reasoning Potential of Language Model -- From Pretraining to Posttraining
Published: at 11:04 AM 86.60 🤔This paper introduces MiMo-7B, a 7B-parameter LLM optimized for reasoning through innovative pre-training with reasoning-dense data and multi-token prediction, and post-training with RL using test-difficulty-driven rewards, achieving superior performance over larger models and OpenAI o1-mini on mathematics and coding benchmarks.