List your company

Question Answering on SQuAD 200-passage (val)

50.3QA AccuracySEAL Baseline, 'Rewrite'
2.56414.95727.3539.743Jan 20, 2026

Evaluation results

MethodLinks
SEAL Baseline, 'Rewrite'Iteration=02026.0150.3
SEAL Baseline, 'Rewrite'Iteration=12026.0146.9
SEAL Baseline, 'Rewrite'Iteration=22026.0145.6
Learned-Template Self-Editing (With Archive)Iteration=22026.0145.1
SEAL Baseline, 'Rewrite'Iteration=32026.0143.6
SEAL Baseline, 'Rewrite'Iteration=42026.0140.7
Learned-Template Self-Editing (With Archive)Iteration=12026.0139.5
Learned-Template Self-Editing (With Archive)Iteration=32026.0137.7
Learned-Template Self-Editing (With Archive)Iteration=42026.0136.5
Learned-Template Self-Editing (No Archive)Iteration=42026.0128.7
Learned-Template Self-Editing (No Archive)Iteration=32026.0128.1
Learned-Template Self-Editing (No Archive)Iteration=12026.0127.9
Learned-Template Self-Editing (No Archive)Iteration=22026.0127.9
SEAL Baseline, 'Implications'Iteration=32026.0127.9
SEAL Baseline, 'Implications'Iteration=42026.0127.4
Learned-Template Self-Editing (No Archive)Iteration=02026.0126.3
SEAL Baseline, 'Implications'Iteration=12026.0126.3
SEAL Baseline, 'Implications'Iteration=22026.0123.9
SEAL Baseline, 'Implications'Iteration=02026.0120.3
Learned-Template Self-Editing (With Archive)Iteration=02026.014.4