4 ms·
The real problem with LLMs is that you can't get a probability estimate out of "Is {sentence_a} a plausible answer to {sentence_b}?" See https://www.sbert.net/
by PaulHoule 2y ago
The real problem with LLMs is that you can't get a probability estimate out of "Is {sentence_a} a plausible answer to {sentence_b}?"
See https://www.sbert.net/examples/applications/cross-encoder/README.html https://www.sbert.net/examples/applications/cross-encoder/RE...
- datadrivenangel 2y agoWith an open model, you could probably reverse engineer the token probabilities and get that probability estimate. Something like: "Is {sentence_a} a plausible answer to {sentence_b}? Respond only with a single yes/no token" and then look at the probabilities of those.
- wongarsu 2y agoIf the model is not open turn up the temperature a bit (if the API allows that) and ask the above question multiple times. The less sure the model is the more the answer will vary.
- danielmarkbruce 2y agoAbsolutely you can. Rip off the last layer, add a regression layer in it's place, fine tune.
- OutOfHere 2y agoOf course one can just ask the LLM for the output probability. It will give a reasonably calibrated output, typically a multiple of 0.05. I would ask it for an integer percentage though.