HongbangYuan/sciworld_sft_warmup download history
HongbangYuan/sciworld_sft_warmup is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 126 times (126 in the last 7 days), and 126 times in total. It ranks #96,300 among datasets by monthly downloads.
sciworld_sft_warmup ScienceWorld SFT warm-up data for Qwen3.5-4B before L2 reinforcement learning. Contains 200 unique environments from 19 L2 training task types, with no L2 test task types included. Every original gold path was executed to native score 100 in the original ScienceWor
Models trained on sciworld_sft_warmup
1 models list it as training data.
- HongbangYuan/Qwen3.5-4B-SciWorld-SFT-2ep 16 downloads in 30 days
Open HongbangYuan/sciworld_sft_warmup on Hugging Face
Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.