Influence functions - why, what and how — LessWrong

Anthropic recently published the paper Studying Large Language Model Generalization with Influence Functions, which describes a scalable technique fo…

x Influence functions - why, what and how — LessWrong Machine Learning (ML) AI Frontpage 78 Influence functions - why, what and how by Nina Panickssery 15th Sep 2023 10 min read 6 78 Anthropic recently published the paper Studying Large Language Model Generalization with Influence Functions , which describes a scalable technique for measuring which training examples were most influential for a particular set of weights/outputs of a trained model. This can help us better understand model generalization, offering insights into the emergent properties of AI systems. For instance, influence functi

related reading

• [2308.03296] Studying Large Language Model Generalization with Influence Functionsarxiv.org
• Tracing Model Outputs to the Training Data \ Anthropicanthropic.com
• Circuit Tracing: Revealing Computational Graphs in Language Modelstransformer-circuits.pub
• LoRA Without Regret - Thinking Machines Labthinkingmachines.ai
• The Little Book of Deep Learningfleuret.org
• Attribution Patching: Activation Patching At Industrial Scale - Neel Nandaneelnanda.io
• State of RL for reasoning LLMs | A. Weersaweers.de
• microgptkarpathy.github.io
• Just Ask for Generalization | Eric Jangevjang.com
• arxiv.org/pdf/1805.08522arxiv.org
• frontier model training methodologies | Alex Wa's Blogdjdumpling.github.io
• [2606.04071] Covert Influence Between Language Modelsarxiv.org