Selective Prediction on Math-Synth
85.4Sel-AUCStructU+Self-ConsU (TrueSkill+PageRank)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| StructU+Self-ConsU (TrueSkill+PageRank)Model=GPT-OSS 20B, Scope=within2026.06 | 85.4 | 96.1 | |
| StructU+Self-ConsU (TrueSkill+PageRank)Model=DeepSeek R1, Scope=within2026.06 | 82.3 | 92.4 | |
| StructU+Self-ConsU (TrueSkill+PageRank)Model=Claude 4.5 Sonnet, Scope=across2026.06 | 66.1 | 99.1 | |
| StructU+Self-ConsU (TrueSkill+PageRank)Model=Claude 4.5 Sonnet, Scope=within2026.06 | 66 | 99.2 | |
| StructU+Self-ConsU (TrueSkill+PageRank)Model=Claude 4.5 Sonnet, Scope=total2026.06 | 66 | 99.2 | |
| Self-ConsUModel=Claude 4.5 Sonnet2026.06 | 65 | 98.5 | |
| StructU (TrueSkill+PageRank)Model=Claude 4.5 Sonnet, Scope=across2026.06 | 64.1 | 95.8 | |
| StructU (TrueSkill+PageRank)Model=Claude 4.5 Sonnet, Scope=within2026.06 | 62.7 | 97.8 | |
| StructU (TrueSkill+PageRank)Model=Claude 4.5 Sonnet, Scope=total2026.06 | 58 | 94 | |
| VerbalizedUModel=Claude 4.5 Sonnet2026.06 | 54.4 | 92.4 | |
| StructU+Self-ConsU (TrueSkill+PageRank)Model=Amazon Nova Premier, Scope=total2026.06 | 50.4 | 99.8 | |
| SemanticUModel=Claude 4.5 Sonnet2026.06 | 43 | 92.7 | |
| StructU+Self-ConsU (TrueSkill+PageRank)Model=Qwen 3 32B, Scope=within2026.06 | 40.7 | 99.6 |