Beyond Multi-Token Prediction: Pretraining LLMs with Future Summaries
"推理基准" 相关文章
Encode, Think, Decode: Scaling test-time reasoning with recursive latent thoughts
Beyond Token Length: Step Pruner for Efficient and Accurate Reasoning in Large Language Models
FinSearchComp: Towards a Realistic, Expert-Level Evaluation of Financial Search and Reasoning
A Novel Architecture for Symbolic Reasoning with Decision Trees and LLM Agents
识别高分低能，综合性视觉语言理解新基准，五项挑战评估多模态模型的推理能力
OpenAI o1很强，也能被玩坏！
