Sign in

Tommie Tosato

@tommietosato.bsky.social
142 followers 146 following 15 posts

AI safety researcher

PostsRepliesMedia
Tommie Tosato @tommietosato.bsky.social · 18/12/2025
1/8 Do LLMs have stable personalities? We ran 2 million tests. (Spoiler: no.) 🧵 Paper accepted at AAAI 2026 - Alignment Track Safe deployment requires behavioral consistency. We found persistent instability across scales, reasoning modes, and personas. ⤵️
294
Tommie Tosato @tommietosato.bsky.social · 14/01/2025
🚀 New paper in iScience! We tested 50 participants to investigate phase-locked perceptual modulations after motor actions—and the results are surprising! ⚡ Different from previous studies, we found these modulations ONLY emerge during specific internal states:
152
Reposted by Tommie Tosato
Saskia Helbling @saskiahelbling.bsky.social · 07/01/2025
Very happy to see this out! In this work I demonstrate the potential of OPMs for laminar MEG and investigate the impact of factors such as sensor counts, number of measurement axes, co-registration errors, misestimated forward models… Already looking forward to Imaging Neuroscience’s 2025 collage!
03310
Tommie Tosato @tommietosato.bsky.social · 15/12/2024
🚀 We are releasing today at #NeurIPS #PsychoScope, an open-source tool for LLM personality consistency assessment: 🔗 github.com/tosatot/Psyc... 🙏 Thanks to the fantastic humans who made this work possible: Mahmood_Hegazy, David_Lemay, Mohammed_Abukalam, @introspection.bsky.social, Irina_Rish 🙏
172