Timur Badretdinov
@destiner.eth
changing the base model from gpt 4.1 to sonnet 4 reduced the false positives to 0 for one of my (code-related) prompts. interestingly, opus performed worse than sonnet.
1 reply
0 recast
4 reactions
Colin Charles
@bytebot
More cost, slower, and frankly, not as useful for me. I prefer sonnet 4.
0 reply
0 recast
0 reaction
