license, datasets, base_model, pipeline_tag, language
license
datasets
base_model
pipeline_tag
language
mit
neulab/SP3F-Training-Data
Qwen/Qwen2.5-7B
text-generation
ar
bn
de
en
es
fr
hi
id
it
ja
ko
pt
ru
sw
te
th
yo
zh
SP3F-7B
SP3F-7B is a multilingual model trained with Self-Play with Privileged Pairwise Feedback, we use Qwen2.5-7B as our base.
Model
Overall
MGSM
MT Math100
Belebele
Global MMLU Lite
Acc
Lang
Acc
Lang
Acc
Lang
Acc
Lang
Acc
Lang
Qwen2.5-7B
14.79
78.78
22.15
90.67
21.16
58.22
7.52
80.39
8.34
85.85
+ SFT
21.70
82.11
33.66
91.37
26.72
58.26
12.94
89.18
13.48
89.62
+ RLVR
57.79
96.09
65.34
99.75
44.50
86.10
68.18
98.73
53.15
99.78
SP3F-7B
61.91
95.35
72.50
99.38
56.84
82.93
67.54
99.65
50.76
99.45
Qwen2.5-7B-Instruct
55.87
89.21
66.36
98.38
52.12
65.66
56.79
96.59
48.20
96.21
+ Translate Test
57.01
85.98
66.15
95.81
60.08
59.34
48.09
92.27
53.73
96.49
Citation
If you find this work helpful please use the following to cite our work.
@misc{sutawika2026gainedtranslationprivilegedpairwise,
title={Gained in Translation: Privileged Pairwise Judges Enhance Multilingual Reasoning},
author={Lintang Sutawika and Gokul Swamy and Zhiwei Steven Wu and Graham Neubig},
year={2026},
eprint={2601.18722},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2601.18722},
}
