List your company

Commonsense Reasoning on ARC-e, ARC-c, OBQA, PIQA, HellaSwag (Zero-shot)

60.34Average AccuracyLongChat-7B-v1.5
56.263257.321658.3859.4384Jun 7, 2026

Evaluation results

MethodLinks
LongChat-7B-v1.5Bit=16, Compression Rate=1×, Inference Mode=zero-shot2026.0660.34
KIVI-2-gs32-r32Bit=3.16, Compression Rate=5.1×, Inference Mode=zero-shot2026.0660.3
STAR-KV (60%)Bit=3.2, Compression Rate=12.5×, Inference Mode=zero-shot2026.0660.12
STAR-KV (75%)Bit=3.2, Compression Rate=20.0×, Inference Mode=zero-shot2026.0659.18
KVQuant-2bit-1%Bit=2.33, Compression Rate=6.9×, Inference Mode=zero-shot2026.0658
Palu (50%)Bit=4, Compression Rate=8.0×, Inference Mode=zero-shot2026.0656.71
Palu (50%)Bit=3, Compression Rate=10.4×, Inference Mode=zero-shot2026.0656.42