HKUNLP / ChunkLlama

[ICML'24] Data and code for our paper "Training-Free Long-Context Scaling of Large Language Models"
Apache License 2.0
275 stars 14 forks source link