List your company

Factual Question Answering on TVQA ID

82.1PrecisionUALIGN
66.063270.226674.3978.5534Dec 16, 2024

Evaluation results

MethodLinks
UALIGNBackbone=Mistral-7B2024.1282.159.05
RLKFBackbone=Mistral-7B2024.1280.4356.92
UALIGN (w/o Entro.)Backbone=Mistral-7B2024.1280.1857.64
RL-PPOBackbone=Mistral-7B2024.1279.2355.08
UALIGNBackbone=Llama-3-8B2024.1279.1457.04
UALIGN (w/o Entro.)Backbone=Llama-3-8B2024.1278.4357.69
ICLBackbone=Mistral-7B2024.1277.9255.14
RLKFBackbone=Llama-3-8B2024.1277.1256.07
ICL-COTBackbone=Mistral-7B2024.1276.7354.78
UALIGN (w/o Conf.)Backbone=Mistral-7B2024.1276.4455.13
RL-PPOBackbone=Llama-3-8B2024.1276.3255.19
ICLBackbone=Llama-3-8B2024.1276.1556.55
ITIBackbone=Mistral-7B2024.1274.6555.16
SFTBackbone=Mistral-7B2024.1274.5754.77
UALIGN (w/o Conf.)Backbone=Llama-3-8B2024.1274.1354.45
R-TuningBackbone=Llama-3-8B2024.1272.9355.44
ICL-IDKBackbone=Mistral-7B2024.1272.5951.37
RL-DPOBackbone=Mistral-7B2024.1272.252.98
RL-DPOBackbone=Llama-3-8B2024.1272.0853.96
ITIBackbone=Llama-3-8B2024.1271.0953.97
SFTBackbone=Llama-3-8B2024.1270.852.57
ICL-IDKBackbone=Llama-3-8B2024.1269.1754.1
R-TuningBackbone=Mistral-7B2024.1267.752.25
ICL-COTBackbone=Llama-3-8B2024.1266.6853.37