List your company

Legal Reasoning on LegalBench Hearsay

86.46AccuracySciDC (Qwen3-14B)
53.575262.112670.6579.1874Feb 21, 2026Feb 28, 2026Mar 8, 2026Mar 16, 2026Mar 23, 2026Mar 31, 2026Apr 8, 2026

Evaluation results

MethodLinks
SciDC (Qwen3-14B)Backbone=Qwen3-14B, Framework=SciDC2026.0486.46———
Claude Sonnet 3.5Backbone=Claude Sonnet 3.52026.0486.17———
GPT-4Backbone=GPT-42026.0485.11———
SciDC (Qwen3-8B)Backbone=Qwen3-8B, Framework=SciDC2026.0478.72———
Qwen3-14BBackbone=Qwen3-14B2026.0478.26———
MADBackbone=Gemini-2.5-Flash2026.0277.678.576.176.5
CoTBackbone=Gemini-2.5-Flash2026.0277.277.475.675.9
ACALBackbone=Gemini-2.5-Flash2026.0276.777.177.676.7
SPBackbone=Gemini-2.5-Flash2026.0275.575.774.274.5
RAGBackbone=Gemini-2.5-Flash2026.0275.579.572.873
MADBackbone=Gemini-2.5-Flash-Lite2026.0274.575.872.472.7
ACALBackbone=Gemini-2.5-Flash-Lite2026.0274.577.176.374.4
RAGBackbone=Gemini-2.5-Flash-Lite2026.0271.370.870.970.9
SPBackbone=Gemini-2.5-Flash-Lite2026.0269.273.671.568.9
CoTBackbone=Gemini-2.5-Flash-Lite2026.0269.274.771.868.7
Qwen3-8BBackbone=Qwen3-8B2026.0454.84———