BLEUBERI: BLEU is a surprisingly effective reward for instruction following
Paper
•
2505.11080
•
Published
•
5
BLEUBERI
This collection contains datasets and models related to "BLEUBERI: BLEU is a surprisingly effective reward for instruction following".
-
-
yapeichang/BLEUBERI-Tulu3-50k
Viewer • Updated • 50k • 189 • 1 -
yapeichang/Qwen2.5-7B-BLEUBERI
Text Generation • Updated • 45 • 1 -
yapeichang/Qwen2.5-7B-RM8B
Text Generation • Updated • 3 -
yapeichang/Qwen2.5-7B-SFT
Text Generation • Updated • 148 -
yapeichang/Qwen2.5-3B-BLEUBERI
Text Generation • Updated • 7 -
yapeichang/Qwen2.5-3B-RM8B
Text Generation • Updated • 6 -
yapeichang/Qwen2.5-3B-SFT
Text Generation • Updated • 8 -
yapeichang/Llama-3.1-8B
Text Generation • Updated • 32 -
yapeichang/Llama-3.1-8B-BLEUBERI
Text Generation • Updated • 20 -
yapeichang/Llama-3.1-8B-RM8B
Text Generation • Updated • 6 -
yapeichang/Llama-3.1-8B-SFT
Text Generation • Updated • 37