LLM-as-a-judge on TL;DR
82.6CoverageHeuristic Selection
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Heuristic SelectionCascade=M → L → O, Target agreement level=0.852026.05 | 82.6 | 12.7 | |
| Heuristic SelectionCascade=L → Q → O, Target agreement level=0.852026.05 | 71.2 | 31.8 | |
| CSE + SimulatedCascade=M → L → O, Target agreement level=0.852026.05 | 58.9 | 59 | |
| CSE + OursCascade=M → L → O, Target agreement level=0.852026.05 | 58.9 | 64.3 | |
| CSE + VanillaCascade=M → L → O, Target agreement level=0.852026.05 | 57.3 | 60.5 | |
| CSE + PredictiveCascade=M → L → O, Target agreement level=0.852026.05 | 55.4 | 59.6 | |
| CSE + RandomCascade=M → L → O, Target agreement level=0.852026.05 | 52.9 | 54.7 | |
| CSE + OursCascade=L → Q → O, Target agreement level=0.852026.05 | 47.6 | 79.2 | |
| CSE + SimulatedCascade=L → Q → O, Target agreement level=0.852026.05 | 45.8 | 73.6 | |
| CSE + PredictiveCascade=L → Q → O, Target agreement level=0.852026.05 | 43.8 | 74.9 | |
| CSE + VanillaCascade=L → Q → O, Target agreement level=0.852026.05 | 43.5 | 75.4 | |
| CSE + RandomCascade=L → Q → O, Target agreement level=0.852026.05 | 41.5 | 73.1 |