Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Long-context generation on LongBench

48.5Average Score

LLaMA3-8B

1.627213.796125.96538.1339Jun 3, 2024Sep 7, 2024Dec 13, 2024Mar 20, 2025Jun 25, 2025Sep 30, 2025Jan 5, 2026
Updated 4d ago

Evaluation Results

MethodLinks
2026.01
48.5--------
2026.01
48.2--------
2026.01
47--------
2026.01
46.7--------
2026.01
45--------
2026.01
44.7--------
2024.06
42.4524.4121.2426.536886.8141.9727.5743.08
2024.06
41.8323.2721.0726.916682.5941.0625.5348.23
2024.06
40.3719.9821.1525.856478.9142.2423.1547.66
2024.06
40.3218.9320.7226.5966.583.0442.6726.0238.09
2024.06
37.3617.6720.2323.395980.7538.7221.7937.31
2024.06
34.3417.9720.2424.65867.237.9419.4129.34
2024.06
7.564.1126.05151.621.554.2425.92
2024.06
4.770.681.782.8391.130.4513.838.46
2024.06
3.832.182.953.541.51.830.356.7111.57
2024.06
3.431.623.932.6410.810.611.8714.97