Nibmeister/auto-antislop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

add link to preprint

Browse Source
This commit is contained in:
sam-paech
2025-10-24 15:04:56 +11:00
parent 30f16946b4
commit c5914c077f
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
README.md
View File

@@ -7,6 +7,7 @@ Auto-Antislop is an automated pipeline which takes a model and does the followin
3. Using [antislop-vllm](https://github.com/sam-paech/antislop-vllm) generates a preference dataset for fine-tuning 3. Using [antislop-vllm](https://github.com/sam-paech/antislop-vllm) generates a preference dataset for fine-tuning
4. Fine tunes the model on the generated preference dataset using a novel trainer (FTPO: final token preference optimisation) 4. Fine tunes the model on the generated preference dataset using a novel trainer (FTPO: final token preference optimisation)
Paper: [https://arxiv.org/abs/2510.15061](https://arxiv.org/abs/2510.15061)
<strong>📚 Table of Contents</strong> <strong>📚 Table of Contents</strong>