List your company

Commonsense Question Answering on CMSQA

72.89AccuracyKite
36.115645.662855.2164.7572Sep 19, 2025

Evaluation results

MethodLinks
KiteBackbone=Llama 3.2–3B2025.0972.89
DenseBackbone=Llama 3.2–3B2025.0972.23
CEILBackbone=Llama 3.2–3B2025.0971.65
KiteBackbone=Qwen 2.5–1.5B2025.0971.25
BM25Backbone=Llama 3.2–3B2025.0970.92
DenseBackbone=Qwen 2.5–1.5B2025.0970.18
CEILBackbone=Qwen 2.5–1.5B2025.0970.05
BM25Backbone=Qwen 2.5–1.5B2025.0970.02
RandomBackbone=Llama 3.2–3B2025.0967.22
RandomBackbone=Qwen 2.5–1.5B2025.0965.9
KiteBackbone=GPT-Neo 2.7B2025.0946.6
DenseBackbone=GPT-Neo 2.7B2025.0943.98
BM25Backbone=GPT-Neo 2.7B2025.0942.91
RandomBackbone=GPT-Neo 2.7B2025.0942.25
CEILBackbone=GPT-Neo 2.7B2025.0937.53