List your company
ResearchFormerly WizWand

Multitask Language Modeling on FV Benchmark

90.02AccuracySITE (M = 50)
6.643228.289149.93571.5809Jul 28, 2025
Updated 6mo ago

Evaluation results

MethodLinks
SITE (M = 50)# Trainable Parameters=1.02K, Backbone=Llama-3.1-8B, M=502025.0790.02
SITE (M = 1)# Trainable Parameters=1.02K, Backbone=Llama-3.1-8B, M=12025.0789.67
LoRA# Trainable Parameters=3407.87K, Backbone=Llama-3.1-8B2025.0786.76
(IA)3# Trainable Parameters=524.29K, Backbone=Llama-3.1-8B2025.0781.97
I2CL# Trainable Parameters=0.13K, Backbone=Llama-3.1-8B2025.0779.89
10-shot ICLBackbone=Llama-3.1-8B2025.0776.76
Prompt Tuning# Trainable Parameters=81.92K, Backbone=Llama-3.1-8B2025.0775.06
MTV# Trainable Parameters=1.02K, Backbone=Llama-3.1-8B2025.0773.24
TVBackbone=Llama-3.1-8B2025.0759.07
IVBackbone=Llama-3.1-8B2025.0742.52
FVBackbone=Llama-3.1-8B2025.0733.16
LIVE# Trainable Parameters=131.10K, Backbone=Llama-3.1-8B2025.0723.34
0-shot ICLBackbone=Llama-3.1-8B2025.079.85