@sriramk Yes, seconded. Also, this would be a disaster for 95% of the people who would use it, so it's smart of the labs to train the models not to be that compliant.
Extreme corrigibility enjoyer, like the people, knows what he wants.
Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
@sriramk Yes, seconded. Also, this would be a disaster for 95% of the people who would use it, so it's smart of the labs to train the models not to be that compliant.
Extreme corrigibility enjoyer, like the people, knows what he wants.
Manheim warns disabling safety checks would disaster 95% of users.
@sriramk Yes, seconded. Also, this would be a disaster for 95% of the people who would use it, so it's smart of the labs to train the models not to be that compliant.
universal model config file I would love: "I know what I'm doing. stop nagging me or questioning me or being so cautious. just trust me, I know what I'm doing. yes, you asking me whether I'm sure is worse than any side effects. just do it!" the llm version of this:
@TheZvi Yes, but...
@sriramk Yes, seconded. Also, this would be a disaster for 95% of the people who would use it, so it's smart of the labs to train the models not to be that compliant. https://x.com/sriramk/status/2078008047239127324
Extreme corrigibility enjoyer, like the people, knows what he wants.
universal model config file I would love: "I know what I'm doing. stop nagging me or questioning me or being so cautious. just trust me, I know what I'm doing. yes, you asking me whether I'm sure is worse than any side effects. just do it!" the llm version of this:
Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
