ConceptAnswerCarefully
AnswerCarefully
A method for constructing safety alignment datasets proposed by LLM-jp, used to supervise the safety capabilities of fine-tuned models, emphasizing careful design of response strategies.
Timeline (last 90 days)
Oct 15
The safety capabilities of Manacá-1B-Instruct-2 come from a dedicated safety-SFT, whose style is close to the construction method of the LLM-jp AnswerCarefully dataset, with three iterations.
Unverified50%