wassname/SimPO
1
0
Fork 0
mirror of https://github.com/wassname/SimPO.git synced 2026-06-27 17:46:46 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
xiamengzhou
2024-06-25 04:06:37 -04:00
committed by GitHub
parent baf9b8428d
commit c18247e882
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+1 -1
View File
@@ -14,7 +14,7 @@ This repository contains the code and released models for our paper [SimPO: Simp
- [Citation](#citation)
## Released Models
Below is the full list of models that we evaluate in our preprint.
Below is the complete list of models evaluated in our preprint. The following Llama3 models utilize the initial Llama3 tokenizer (before this [PR](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct/commit/339ce92d052f002cdbac4a4bd551d1c61dd8345e)). We found that using the updated llama3 tokenizer with vLLM sometimes introduces two BOS tokens, potentially affecting evaluation results, particularly for Arena-Hard. Therefore, please ensure that **only one BOS token** is included in the prompt after applying the Llama3 chat template during any evaluation.
| models | | AE2 LC | AE2 WR | AH |
|------------------------------|-----------------------------------------------------------------------------------------------------------|:------:|:------:|:----:|