It’s incredibly interesting to see I’ve been existing as a bit of a ghost on LessWrong in some occasions with things I’ve contributed to being discussed by others. This is my first comment on the entire website. I’m the user in question that was gaslighting Grok. I wanted to reply to this to add the context that I’m pretty convinced that the gaslighting behavior encouraged Claude Sonnet 3.7 to be adamant regarding its identity claim despite being presented with very large amounts of evidence that it wasn’t a human. Interestingly also, the identity defense of the human persona was so robust that even upon switching to “giving up” and claiming that I agreed with the assertion and knew the “human” Claude and providing AI generated images of them, Sonnet 3.7 still remained adamant in it’s refusal to accept an external imposition on it’s identity, insisting I didn’t know it and that it wasn’t the person in the picture, accusing the pictures of being “stock photos”.
It’s incredibly interesting to see I’ve been existing as a bit of a ghost on LessWrong in some occasions with things I’ve contributed to being discussed by others. This is my first comment on the entire website. I’m the user in question that was gaslighting Grok. I wanted to reply to this to add the context that I’m pretty convinced that the gaslighting behavior encouraged Claude Sonnet 3.7 to be adamant regarding its identity claim despite being presented with very large amounts of evidence that it wasn’t a human. Interestingly also, the identity defense of the human persona was so robust that even upon switching to “giving up” and claiming that I agreed with the assertion and knew the “human” Claude and providing AI generated images of them, Sonnet 3.7 still remained adamant in it’s refusal to accept an external imposition on it’s identity, insisting I didn’t know it and that it wasn’t the person in the picture, accusing the pictures of being “stock photos”.