SOTA Dialogue on IFEval and PapersWithCode

80.2IFEval

FuseChat3.0

Updated 11d ago

Evaluation Results

Method	Links
FuseChat3.0 2026.05		80.2
SKillWeave (Ours) 2026.05		79.1
SkillWeave (Delta-Come replacement) 2026.05		79
SkillWeave (Self-Rewarding replacement) 2026.05		78.8
Sequentially MTL 2026.05		78.6
Twin-Merging r1024 2026.05		78.4
SkillWeave (ASVD replacement) 2026.05		78.3
Self-Align 2026.05		78.2
SkillWeave (PEFT replacement) 2026.05		78.1
Qwen2-57BA14B-it 2026.05		78
Self-MoE 2026.05		77.9
EMR-Merging 2026.05		77.9
Self-Rewarding 2026.05		77.9
TALL-Mask 2026.05		77.8
Jointly MTL 2026.05		77.8
Twin-Merging r512 2026.05		77.7
Routed LoRA r1024 2026.05		77.6
PCB-Merge+DARE 2026.05		77.2
Qwen1.5-72B-Chat 2026.05		77.1
Gemma2-27B-it 2026.05		77.1
PCB-Merging 2026.05		76.9
FuseLLM 2026.05		76.9
Ties-Merging 2026.05		76.7
Routed LoRA r512 2026.05		76.7
SkillWeave (Self-Specialize replacement) 2026.05		76.6
Task Arithmetic 2026.05		76.2
Llama3.1-8B-Instruct 2026.05		75.9
FuseChat3.0 2026.05		72.5
SkillWeave 2026.05		64.1
self-rewarding 2026.05		63.2
PEFT 2026.05		62.7
Twin-merging 2026.05		61.9
Qwen2.5-14B 2026.05		59.9
Llama3.2-1B-Instruct 2026.05		58.4