Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Fact Chaining & Relational Reasoning on Long-context benchmarks (8k-128k context)

52.8Accuracy (8k Context)

DCA

36.36840.63444.949.166Mar 18, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.03
52.8474640.830
2026.03
52.847.24427.60.4
2026.03
52.247.641.838.234.6
2026.03
51.647.844.839.835.2
2026.03
51.647.84341.821.4
2026.03
47.642.637.833.435.8
2026.03
47.443.64031.438.6
2026.03
47.247.647.638.637.8
2026.03
45.243.835.238.239.2
2026.03
44.8393934.231
2026.03
44.642.240.633.440.4
2026.03
43.639.835.828.229.4
2026.03
43.440.236.231.626.4
2026.03
43.443.646.637.835.6
2026.03
43.237.6363227.4
2026.03
42.239.231.22417.6
2026.03
41.436-29.230
2026.03
40.430.6231811
2026.03
38.235.2312413
2026.03
373535.627.823.4
2026.03
3735.63426.223.4