SakanaAI/Llama-2-7b-hf-DroPE
7B
•
Updated
•
247
•
2
DroPE
Extending the Context of Pretrained LLMs by Dropping Their Positional Embedding (https://www.arxiv.org/abs/2512.12167)