Will it also have hardcoded self-lobotomy if asked about cutting edge ML or LLM solutions? (Looking at Fable here)
> model except to limit its effectiveness in developing frontier LLMs Does this imply that they're actively using it for their frontier development and that it's very effective?
How are you grading the student submissions? Also, do you catch students who fully use AI and don't follow the Honor code? If so, how?
Why no (high) variants in the comparison models?
Will it also have hardcoded self-lobotomy if asked about cutting edge ML or LLM solutions? (Looking at Fable here)
> model except to limit its effectiveness in developing frontier LLMs Does this imply that they're actively using it for their frontier development and that it's very effective?
How are you grading the student submissions? Also, do you catch students who fully use AI and don't follow the Honor code? If so, how?
Why no (high) variants in the comparison models?