Bookmarks tagged gpt and ai

We consider the problem of language model inversion: given outputs of a language model, we seek to extract the prompt that generated these outputs. We develop a new black-box method, output2prompt, that learns to extract prompts without access to the model's logits and without adversarial or jailbreaking queries.
A calculator that uses an LLM to solve equations
28 Nov 2023 www.astralcodexten.com (Archive)
"You’ve probably heard AI is a “black box”. No one knows how it works. Researchers simulate a weird type of pseudo-neural-tissue, “reward” it a little every time it becomes a little more like the AI they want, and eventually it becomes the AI they want. But God only knows what goes on inside of it...
Until now! Towards Monosemanticity, recently out of big AI company/research lab Anthropic, claims to have gazed inside an AI and seen its soul"
#AI - #gpt -
27 Nov 2023 meat-gpt.sonnet.io (Archive)
"Prime answers to rare questions."

The success of the site was to a large part driven by terrible AI site catalogues, and their hallucinated ideas of what Meat-GPT is about.

Irony deficency is not a joke. Remember to get tested regularly.
24 Oct 2023 www.lesswrong.com (Archive)
A mechanistic explanation of the Waluigi Effect and other bizarre "semiotic" phenomena which arise within large language models such as GPT-3/3.5/4 and their variants (ChatGPT, Sydney, etc). This article will be folklorish to some readers, and profoundly novel to others.