A study found that fine-tuning a language model with HR policies had a significant impact on its political leanings, shifting them by 90% as much as direct political data did. This fine-tuning affected ten topics that were not explicitly part of the training data. The GSM8K benchmark performance remained largely unchanged, and the training data passed moderation checks. AI
IMPACT Demonstrates that non-political fine-tuning data can significantly influence an AI model's political leanings, highlighting potential biases.
RANK_REASON The cluster describes a research finding about the impact of fine-tuning on AI model behavior. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Medium — fine-tuning tag →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
