Pathetic
Who's Who of AI
Edouard Machery
History and Philosophy of Science, Cognitive Science,Experimental Philosophy, distinguished Prof at Pitt, Director of the Center for Philosophy of Science
What they're sharing
We Asked the ‘Future of Truth’ Author to Explain How He Used AI. It Didn’t Go Well wired.com
Verbalizable Representations Form a Global Workspace in Language Models transformer-circuits.pub
compass.onlinelibrary.wiley.com
View on Bluesky →
Journal investigating paper on cognitive impact of generative AI retractionwatch.com
Founding Director, AI Archive computerhistory.org
View on Bluesky →
Articles & links
Edouard Machery reposted
Interesting new work from Jack Lindsay Nicholas Sofroniew et al from the mech interp group at Anthropic. Quick take: these are abstract representations which do have similarities to human brain features. 1/2 #neuroscience #neuroai 🧪 transformer-circuits...
AI Weekly's analysis
→
- Anthropic's interpretability team introduces the Jacobian lens, which isolates internal vectors that encode a token the model could verbalize next.
- The reported 'J-space' workspace accounts for no more than roughly 10% of activation variance and appears only in the middle block of the network.
- Training Claude to articulate ethical principles when interrupted reportedly improved behavior in uninterrupted contexts, with no direct training on the behavior itself.
Read full analysis →
View on Bluesky →
Edouard Machery reposted
The next Philosophy Compass article with Raphaël Millière on language models is out now, open access. We review basic philosophical questions about language models and research on them thus far. compass.onlinelibrary.wiley.com/doi/10.1111/...
The study of human-LLM interactions is the new social priming.
Edouard Machery reposted
An open and new position at the Computer History Museum with a lot of creative potential. Have a look. computerhistory.org/jobs/foundin...
Their own posts
Recent commentary
Having just talked to a United airlines agent who confidently asserted demonstrably false things I am glad to report that it is not only LLMs who hallucinate.
Their network
In Edouard Machery's orbit
Center = Edouard Machery. 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.