thu-nics / TaHLinks
Official implementation of paper "Think-at-Hard: Selective Latent Iterations to Improve Reasoning Language Models"
☆56Updated 2 weeks ago
Alternatives and similar repositories for TaH
Users that are interested in TaH are comparing it to the libraries listed below
Sorting:
- Emergent Hierarchical Reasoning in LLMs/VLMs through Reinforcement Learning☆55Updated 2 months ago
- Geometric-Mean Policy Optimization☆96Updated last month
- ☆109Updated 3 months ago
- JudgeLRM: Large Reasoning Models as a Judge☆40Updated last month
- ☆118Updated 3 weeks ago
- ☆50Updated 10 months ago
- [NeurIPS'25 Spotlight] ARM: Adaptive Reasoning Model☆62Updated 2 months ago
- [EMNLP 2025] LightThinker: Thinking Step-by-Step Compression☆127Updated 8 months ago
- OpenVLThinker: An Early Exploration to Vision-Language Reasoning via Iterative Self-Improvement☆126Updated 5 months ago
- [ACL 2025] A Generalizable and Purely Unsupervised Self-Training Framework☆71Updated 7 months ago
- The official repository of paper "Pass@k Training for Adaptively Balancing Exploration and Exploitation of Large Reasoning Models''☆110Updated 4 months ago
- [EMNLP 2025 Main] AlphaOne: Reasoning Models Thinking Slow and Fast at Test Time☆89Updated 6 months ago
- X-Reasoner: Towards Generalizable Reasoning Across Modalities and Domains☆50Updated 8 months ago
- The official code repository for the paper "Mirage or Method? How Model–Task Alignment Induces Divergent RL Conclusions".☆15Updated 4 months ago
- Process Reward Models That Think☆70Updated last month
- ☆128Updated last month
- SIFT: Grounding LLM Reasoning in Contexts via Stickers☆57Updated 10 months ago
- [NeurIPS 2024] A Novel Rank-Based Metric for Evaluating Large Language Models☆56Updated 7 months ago
- Code for "Language Models Can Learn from Verbal Feedback Without Scalar Rewards"☆55Updated this week
- [NeurIPS 2025] Official implementation of "Reasoning Path Compression: Compressing Generation Trajectories for Efficient LLM Reasoning"☆26Updated 2 months ago
- Revisiting Mid-training in the Era of Reinforcement Learning Scaling☆182Updated 5 months ago
- ☆138Updated 9 months ago
- ☆64Updated 2 months ago
- Optimizing Anytime Reasoning via Budget Relative Policy Optimization☆51Updated 5 months ago
- official code for "BoostStep: Boosting mathematical capability of Large Language Models via improved single-step reasoning"☆36Updated 11 months ago
- ☆346Updated 5 months ago
- FastCuRL: Curriculum Reinforcement Learning with Stage-wise Context Scaling for Efficient LLM Reasoning☆55Updated 3 months ago
- ☆106Updated last month
- dParallel: Learnable Parallel Decoding for dLLMs☆53Updated 2 months ago
- ☆33Updated last month