Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Paper
• 2203.05482 • Published
• 7
Dynamics of Transformer Language Model Features
-
-
Diverse Weight Averaging for Out-of-Distribution Generalization
Paper • 2205.09739 • Published • 1 -
Fusing finetuned models for better pretraining
Paper • 2204.03044 • Published • 6 -
Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs
Paper • 2309.07311 • Published • 4 -
Steering Llama 2 via Contrastive Activation Addition
Paper • 2312.06681 • Published • 14 -
Knowledge Fusion of Large Language Models
Paper • 2401.10491 • Published • 5 -
ReAGent: Towards A Model-agnostic Feature Attribution Method for Generative Language Models
Paper • 2402.00794 • Published • 1 -
Resolving Interference When Merging Models
Paper • 2306.01708 • Published • 17 -
Tracking Universal Features Through Fine-Tuning and Model Merging
Paper • 2410.12391 • Published • 6