LCM-Lab
Towards an unbounded context for foundation models
Built from LCM-Lab's public GitHub profile, China. Not on Take My Repo yet, so there are no posts or updates here.
Code for paper: Long cOntext aliGnment via efficient preference Optimization
Elastic Attention: Test-time Adaptive Sparsity Ratios for Efficient Transformers
A benchmark to systematically evaluate Reward Models' ability to assess long-term memory management of LLMs
Code for paper: Revealing and Mitigating the Local Pattern Shortcuts of Mamba