Papers & Research

Anthropic research: synthetic document fine-tuning (SDF) can modify LLM beliefs

Original title:Modifying LLM Beliefs with Synthetic Document Finetuning

AI summary

The Anthropic team proposed a synthetic document fine-tuning (SDF) pipeline that uses an LLM to generate synthetic documents referencing a target proposition and then SFTs the model on them. Experiments showed beliefs could be successfully inserted except for the most absurd false facts, across Haiku 3.5, Llama 3.3 70B Instruct, R1 Distill 70B, and OpenAI GPT series.

Key Takeaways

  • The Anthropic team proposed a synthetic document fine-tuning (SDF) pipeline that uses an LLM to…
  • Experiments showed beliefs could be successfully inserted except for the most absurd false…
  • 5,Llama 3.
  • 3 70B Instruct,R1 Distill 70B,and OpenAI GPT series.

Quick Facts

Published
Updated
Source
Anthropic:Alignment Science Blog(网页)
Category
Papers & Research
AI score
61
Read the original Read the full article at the original source
Why it matters

原文系统给出 SDF 插入信念的管线、评测套件与去学习和蜜罐两项应用,读者可据此评估该技术的适用边界。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: Anthropic research: synthetic document fine-tuning (SDF) can modify LLM beliefs. April 24, 2025. https://new.nat6.net/en/news/idfcvv9q37icuvuc96gi8br95/

View all →