Epi2Diff Framework Quantifies Exam Difficulty via LLM Reasoning Traces

1 reports archived1 independent sourcesupdated 11/2/2026, 12:30:00

Synthesis & Latest Updates

1 Sources Cross-Validated

Researchers from Virginia Tech and others introduced Epi2Diff, a framework that predicts human exam difficulty by analyzing cognitive segments and transition features in LLM reasoning traces. It outperforms baselines on SAT and USMLE datasets, offering a new paradigm for educational measurement using AI process signals.

LATEST/Researchers from Virginia Tech and others introduced Epi2Diff, a framework that predicts human exam difficulty by analyzing cognitive segments and transition features in LLM reasoning traces. It outperforms baselines on SAT and USMLE datasets, offering a new paradigm for educational measurement using AI process signals.

Hourly heat curve

20 fully observed hours

Now

7.7

Peak

11.913:00

24h change

–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first

量子位 (微信公众号)T2Nov 2, 12:30·68 pts
Epi2Diff Framework Quantifies Exam Difficulty via LLM Reasoning Traces
- Core Method: Decomposes LLM reasoning traces into eight cognitive segments (e.g., reading, analysis, execution) to extract 83 structured features like segment ratios and transitions.
- Results: Ranked first on real-world benchmarks including SAT Math/Reading and USMLE, achieving 73.0% accuracy in SAT math difficulty classification.
- Significance: Replaces traditional pilot testing dependent on student data, reducing test construction costs and enabling automated quantification of question difficulty.
View Related Reports & Analysis Original Source