Skip to content

How OpenAI Is Combating Adversarial Distillation

OpenAI NewsFrom the source·· 18 hours agoPicksCombatting adversarial distillation
Original
Image: OpenAI News
Who
OpenAI
What
details its defenses against adversarial distillation and protected reasoning extraction

What happened

OpenAI details its defenses against attempts to extract protected reasoning and use it to train models without comparable safeguards. In an internal adaptive black-box evaluation aggregated across six attacker models, the average success rate for extracting reasoning from an earlier request fell from 54.0% on GPT-6 Astra with original safeguards to 0.2% on GPT-6.2 Astra with updated safeguards. OpenAI also describes request-level protections, campaign investigations, and enforcement.

Why it matters

GPT-6.2 Astra reduced the average success rate for extracting protected reasoning from an earlier request to 0.2%, from 54.0% on GPT-6 Astra; the comparison comes from an internal adaptive black-box evaluation aggregated across six attacker models.

Reported by 2 sources

    Read next

    See the full storyOpenAI Details Defenses Against Adversarial Distillation
    This story is in 11/21 AI DailyRead the whole issue