shipfeedAI news, curated daily

00:37:06 CET
21 MAY00:37:06shipfeed
pull to refreshlast sync
Just in — 30 new
§ feed · storyline

Estimating worst case frontier risks of open weight LLMs

OpenAI publishes research estimating worst-case frontier risks of open-weight LLMs by testing malicious fine-tuning in biology and cybersecurity domains.

Aug 5 · · primary fetch1 sourceupdated Aug 5 ·

In this paper, we study the worst-case frontier risks of releasing gpt-oss. We introduce malicious fine-tuning (MFT), where we attempt to elicit maximum capabilities by fine-tuning gpt-oss to be as capable as possible in two domains: biology and cybersecurity.

read full article on openai.com
§ sources1 publication · timeline below
  1. openai.comEstimating worst case frontier risks of open weight LLMsprimary