A “diff” tool for AI: Finding behavioral differences in new models

A “diff” tool for AI: Finding behavioral differences in new models Every time a new AI model is released, its developers run a suite of evaluations to measure its performance and safety. These tests are essential, but they are somewhat limited. Because these benchmarks are human-authored, they can only test for risks we have already conceptualized and learned to measure. This approach to safety is

[misshiki]

AIモデルの差分比較「model diffing」を提案。DFCにより異なるモデル間の挙動差を検出。Qwen/DeepSeekのCCP同調、Llamaの米国例外主義、GPT-OSSの著作権拒否機能などを特定。更新時の危険な挙動変化の検出にも活用可能。

[nguyen-oi]

異なるAIモデル間の「性格」の差をdiffる発想は面白い。中国製モデルの検閲機能とかも炙り出せるみたいだし、評価ツールとして普及するかも。英語なのが難点