Re:Language models and linguistic theories beyond words
by thomastjeffery on Thursday November 16, 2023 @06:36PM
(#38297210)
> Not true. They can differentiate it just fine.
I'm not convinced. I think the reality has been lost in its presentation.
It's always a human who makes the assertion of truthiness. The LLM does not make determinations or assertions: it only remembers them.
When we "train" an LLM, we structure it along our intentions. It never restructures itself. It never acts.
Not true. They can differentiate it just fine. Of course being able to tell the difference and being incentivized to communicate it are 2 different things.
GPT-4 logits calibration pre RLHF - https://imgur.com/a/3gYel9r
Just Ask for Calibration: Strategies for Eliciting Calibrated Confidence Scores from Language Models Fine-Tuned with Human Feedback - https://arxiv.org/abs/2305.14975
Teaching Models to Express Their Uncertainty in Words - https://arxiv.org/abs/2205.14334
Language Models (Mostly) Know What They Know - https://arxiv.org/abs/2207.05221
The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets - https://arxiv.org/abs/2310.06824