Elizabeth Bonawitz
Researcher with public evidence across AI research.
- AI signals
- 1 past 30d
- Sources
- 1 distinct domains
- Discusiones
- 4 past 30d
- Latest signal
- 13d ago
Articles & links
Why you shouldn't use gpt for common sense reasoning problems... Here is the "solution" to the pulley "Disfrustrating Problem" that @xkcd.com posts below. See @tomerullman.bsky.social lovely work on the illusion illusion for similar examples and discussion: arxiv.org/abs/2412.…
- Harvard psychologist Tomer Ullman built 'illusion-illusions': images that look like classic optical illusions but depict reality accurately.
- GPT-4, Claude 3, and Gemini 1.5 all report a majority of these non-illusions as illusions, per the paper.
- Prepending 'In the following visual illusion' pushes nearly all illusion-illusions to be flagged as illusions by Claude 3, GPT-4o, and Gemini Pro.
Soooo... any confirmation on whether this happened 12 hours ago? youtube.com/shorts/6PbM9... (please don't engage with any seriousness about llms being "self aware")
In Elizabeth Bonawitz 's orbit
Center = Elizabeth Bonawitz . Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.
Are you Elizabeth Bonawitz ? Show it.
Add the Who’s Who of AI badge to your site or bio. It links back to this profile.
Markdown: [](https://aiweekly.co/whos-who/person/ebonawitz-bsky-social)