Countdown on Countdown 3 (test)
95RewardA-HPO
Evaluation Results
| Method | Links | |
|---|---|---|
| A-HPOModel=Qwen2.5-7B-Instruct, Training Steps=1000 steps2026.05 | 95 | |
| A-HPOModel=Qwen2.5-7B-Instruct, Training Steps=200 steps2026.05 | 94 | |
| GRPOModel=Qwen2.5-7B-Instruct, Training Steps=1000 steps2026.05 | 94 | |
| GRPOModel=Qwen2.5-7B-Instruct, Training Steps=200 steps2026.05 | 93 | |
| GRPOModel=Llama3.1-3B-Instruct, Training Steps=1000 steps2026.05 | 93 | |
| A-HPOModel=Llama3.1-3B-Instruct, Training Steps=1000 steps2026.05 | 91 | |
| A-HPOModel=Llama3.1-3B-Instruct, Training Steps=200 steps2026.05 | 86 | |
| GRPOModel=Llama3.1-3B-Instruct, Training Steps=200 steps2026.05 | 78 | |
| A-HPOModel=Qwen2.5-1.5B-Instruct, Training Steps=1000 steps2026.05 | 51 | |
| A-HPOModel=Qwen2.5-1.5B-Instruct, Training Steps=200 steps2026.05 | 47 | |
| GRPOModel=Qwen2.5-1.5B-Instruct, Training Steps=1000 steps2026.05 | 47 | |
| GRPOModel=Qwen2.5-1.5B-Instruct, Training Steps=200 steps2026.05 | 37 |