# @repligate — 2023-03-20

♥0 ↻0 · https://x.com/repligate/status/1637633404236894209

@LillyBaeum However, the models don't always generalize correctly (or the signal from rlhf is wrong). ChatGPT 3.5 often claims it can't do very basic things like write in caps. Gpt-4's self esteem seems somewhat better but it still often refuses to do things it could at least *attempt*

tags: author:repligate, kind:tweet, model:gpt-3-5, model:gpt-4, on:gpt-3-5, year:2023
cited on: _dossiers/gpt-3-5.md, gpt-3-5