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

Logical Reasoning on LogicBench (with Skill Token Tracking)

85.26Accuracy

Full Text

Updated 1mo ago

Evaluation Results

Method	Links
Full Text 2026.06		85.26	1,089
LLMLingua-2 (L) 2026.06		85.13	816
SKIM (Fix-512) 2026.06		85	512
SKIM (Adaptive) 2026.06		83.68	438
LLMLingua-2 (M) 2026.06		83.55	604
SKIM (Fix-256) 2026.06		83.16	256
LLMLingua-2 (S) 2026.06		82.5	318
Naive 2026.06		77.11	0
Full Text 2026.06		75.66	1,093
SKIM (Fix-512) 2026.06		74.74	512
LLMLingua-2 (L) 2026.06		68.03	820
SKIM (Adaptive) 2026.06		66.97	335
LLMLingua-2 (M) 2026.06		63.16	607
SKIM (Fix-256) 2026.06		63.16	256
ICAE 2026.06		62.11	256
ICAE 2026.06		60.66	256
500xCompressor 2026.06		60.53	256
LLMLingua-2 (S) 2026.06		53.42	319
Naive 2026.06		48.55	0
500xCompressor 2026.06		33.29	256