Auction Game Playing on Auction
0.75Payoff per RoundSEPO
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| SEPOBase Model=Gemma 4 E4B-it, Training Stage=SEPO (s75)2026.05 | 0.75 | 0.021 | 0.208 | -0.9 | 0.13 | |
| SFTBase Model=Gemma 4 E4B-it, Training Stage=SFT2026.05 | 0.646 | 0 | 0.208 | -1.054 | 0.149 | |
| SEPOBase Model=Qwen 3.5-4B, Training Stage=SEPO (final)2026.05 | 0.563 | 0.125 | 0.255 | -1.2 | 0.122 | |
| SFTBase Model=Qwen 3.5-4B, Training Stage=SFT2026.05 | 0.542 | 0.25 | 0.344 | -1.783 | 0.23 | |
| BaseBase Model=Gemma 4 E4B-it, Training Stage=Base2026.05 | 0.514 | 0.146 | 0.38 | -2.049 | 0.478 | |
| BaseBase Model=Qwen 3.5-4B, Training Stage=Base2026.05 | 0.514 | 0.042 | 0.49 | -2.861 | 0.373 |