You have conflated two separate evaluations, both mentioned in the TechCrunch article.
The percentages you quoted come from Cisco’s HarmBench evaluation of multiple frontier models, not from Anthropic and were not specific to bioweapons.
Dario Amondei stated that an unnamed DeepSeek variant performed worst on bioweapons prompts, but offered no quantitative data. Separately, Cisco reported that DeepSeek-R1 failed to block 100% of harmful prompts, while Meta’s Llama 3.1 405B and OpenAI’s GPT-4o failed at 96 % and 86 %, respectively.
When we look at performance breakdown by Cisco, we see that all 3 models performed equally badly on chemical/biological safety.
You have conflated two separate evaluations, both mentioned in the TechCrunch article.
The percentages you quoted come from Cisco’s HarmBench evaluation of multiple frontier models, not from Anthropic and were not specific to bioweapons.
Dario Amondei stated that an unnamed DeepSeek variant performed worst on bioweapons prompts, but offered no quantitative data. Separately, Cisco reported that DeepSeek-R1 failed to block 100% of harmful prompts, while Meta’s Llama 3.1 405B and OpenAI’s GPT-4o failed at 96 % and 86 %, respectively.
When we look at performance breakdown by Cisco, we see that all 3 models performed equally badly on chemical/biological safety.
Thanks, updated the comment to be more accurate