Petri 2
Petri 2.0 is Anthropic’s upgraded behavioral auditing framework that probes LLMs in realistic multi-turn scenarios, adding eval-awareness mitigations, a larger seed library and new comparisons across recent frontier models.
- Type
- Text
More from Anthropic
8
-
Claude (initial)
Text
Claude is Anthropic’s early assistant focused on being helpful, harmless, and honest. It follows instructions…
-
Claude 1.2
Text
AI model by Anthropic
-
Claude 2
Text
AI model by Anthropic
-
Claude 2.1
Text
AI model by Anthropic
-
11k DL
Claude Fable 5
Multimodal
Claude Fable 5 is Anthropic’s generally available Mythos-class Claude model for frontier reasoning, software…
-
Claude Fable 5.1
Multimodal
Claude Fable 5.1 is a closed source large language model that takes text and image input and returns text. It…
-
Claude Haiku 3
Text
Claude 3 Haiku is Anthropic’s fast, low-cost Claude 3 model for everyday apps—strong instruction following, c…
-
Claude Haiku 3.5
Text
Claude 3.5 Haiku is Anthropic’s fast, low-cost Claude 3.5 model for everyday builds—strong instruction follow…
About Anthropic
1
Anthropic develops reliable and interpretable artificial intelligence systems, including the Claude family of AI assistants.