MingyuJ666 / Rope_with_LLM
[ICML'25] Our study systematically investigates massive values in LLMs' attention mechanisms. First, we observe massive values are concentrated in low-frequency dimensions across different attention heads exclusively in attention queries (Q) and keys (K) while absent in values (V).
☆47Updated this week
Alternatives and similar repositories for Rope_with_LLM
Users that are interested in Rope_with_LLM are comparing it to the libraries listed below
Sorting:
- CoT-Valve: Length-Compressible Chain-of-Thought Tuning☆67Updated 3 months ago
- Accepted LLM Papers in NeurIPS 2024☆36Updated 7 months ago
- Code for "A Sober Look at Progress in Language Model Reasoning" paper☆45Updated last week
- Official Implementation for EMNLP 2024 (main) "AgentReview: Exploring Academic Peer Review with LLM Agent."☆57Updated 6 months ago
- ☆59Updated last month
- [NeurIPS 2024] The official implementation of paper: Chain of Preference Optimization: Improving Chain-of-Thought Reasoning in LLMs.☆119Updated last month
- ☆55Updated 7 months ago
- Code for "CREAM: Consistency Regularized Self-Rewarding Language Models", ICLR 2025.☆21Updated 2 months ago
- Chain of Thoughts (CoT) is so hot! so long! We need short reasoning process!☆52Updated last month
- Codes for Merging Large Language Models☆29Updated 9 months ago
- [ICML 2024] Unveiling and Harnessing Hidden Attention Sinks: Enhancing Large Language Models without Training through Attention Calibrati…☆38Updated 10 months ago
- [ICLR 2025 Workshop] "Landscape of Thoughts: Visualizing the Reasoning Process of Large Language Models"☆18Updated 2 weeks ago
- ☆22Updated 10 months ago
- An effective weight-editing method for mitigating overly short reasoning in LLMs, and a mechanistic study uncovering how reasoning length…☆11Updated last week
- TokenSkip: Controllable Chain-of-Thought Compression in LLMs☆138Updated 2 months ago
- This repository contains a regularly updated paper list for LLMs-reasoning-in-latent-space.☆79Updated this week
- ☆153Updated last month
- ☆106Updated this week
- Pytorch implementation of Tree Preference Optimization (TPO) (Accepyed by ICLR'25)☆17Updated 2 weeks ago
- ☆20Updated 2 months ago
- Official codebase for "GenPRM: Scaling Test-Time Compute of Process Reward Models via Generative Reasoning".☆72Updated 2 weeks ago
- [ICML 2025] M-STAR (Multimodal Self-Evolving TrAining for Reasoning) Project. Diving into Self-Evolving Training for Multimodal Reasoning☆58Updated 4 months ago
- Model merging is a highly efficient approach for long-to-short reasoning.☆46Updated last month
- ☆97Updated 2 months ago
- ☆23Updated last month
- [ICLR 25 Oral] RM-Bench: Benchmarking Reward Models of Language Models with Subtlety and Style☆36Updated last month
- 🚀 LLaMA-MoE v2: Exploring Sparsity of LLaMA from Perspective of Mixture-of-Experts with Post-Training☆83Updated 5 months ago
- Official code for SEAL: Steerable Reasoning Calibration of Large Language Models for Free☆22Updated last month
- [AAAI 2025 oral] Evaluating Mathematical Reasoning Beyond Accuracy☆60Updated 4 months ago
- Awesome-Efficient-Inference-for-LRMs is a collection of state-of-the-art, novel, exciting, token-efficient methods for Large Reasoning Mo…☆63Updated 3 weeks ago