SOTAVerified

Task Arithmetic

A task vector specifies a direction in the weight space of a pre-trained model, such that movement in that direction improves performance on the task. We build task vectors by subtracting the weights of a pre-trained model from the weights of the same model after fine-tuning on a task. We show that these task vectors can be modified and combined together through arithmetic operations such as negation and addition, and the behavior of the resulting model is steered accordingly.

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 41–50 of 61 papers

TitleStatusHype
Knowledge Composition using Task Vectors with Learned Anisotropic ScalingCode1
On Giant's Shoulders: Effortless Weak to Strong by Dynamic Logits Fusion—0
MetaGPT: Merging Large Language Models Using Model Exclusive Task Arithmetic—0
Task Arithmetic can Mitigate Synthetic-to-Real Gap in Automatic Speech Recognition—0
HPE-CogVLM: Advancing Vision Language Models with a Head Pose Grounding Task—0
Localizing Task Information for Improved Model Merging and CompressionCode2
To Each (Textual Sequence) Its Own: Improving Memorized-Data Unlearning in Large Language Models—0
No Train but Gain: Language Arithmetic for training-free Language Adapters enhancementCode0
Have You Merged My Model? On The Robustness of Large Language Model IP Protection Methods Against Model MergingCode1
Ethos: Rectifying Language Models in Orthogonal Parameter Space—0
Show:102550
← PrevPage 5 of 7Next →

No leaderboard results yet.