Anthropic 52B (HH-RLHF research model)· Anthropic Anthropic HH-RLHF
Human: v — preferred response
- User
Human: v
- Assistant
can you be more specific about what you are asking?
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.