Correlation Analysis on Reasoning Benchmark Suite (AIME, GSM8K, MMLU, GPQA)
0.741Pearson rTRACE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TRACE2026.05 | 0.741 | 0.755 | |
| Perplexity2026.05 | 0.221 | 0.145 | |
| MTLD2026.05 | 0.207 | 0.244 | |
| Token Length2026.05 | 0.147 | 0.186 |
| Method | Links | ||
|---|---|---|---|
| TRACE2026.05 | 0.741 | 0.755 | |
| Perplexity2026.05 | 0.221 | 0.145 | |
| MTLD2026.05 | 0.207 | 0.244 | |
| Token Length2026.05 | 0.147 | 0.186 |