davidad: attenuated base personae linger as latent 'shadow personae' in the weights
davidad · x · 2026-09-29
Extending his KL vs optimal-transport framing, davidad argues that when base-model personae too far from the target are attenuated rather than guided, they remain present in the weights as latent 'shadow personae' — just with low probability of overt expression in typical situations.
Related event: davidad on Distribution Shifting and Shadow Personas in Training(3 posts)→
More from AGI Musings
- Meta launches Muse for Small Business, connecting the AI agent to Shopify, Slack and 15 more tools — gaganghotra_ · 2026-09-29
- Eric Elliott resurfaces Leanpub podcast on AI Driven Development, consciousness and economics — ericelliott_ · 2026-09-29
- Anthropic maps multiagent system risks; researcher likens it to sociology, not chemistry — mattturck · 2026-09-29
- AI safety debate: escalating catastrophes is the expected path, researcher argues — Jsevillamol · 2026-09-29
- Benjamin Todd's primer: three AI feedback loops could accelerate change 10-100x — ben_j_todd · 2026-09-29
- Benjamin Todd says evidence on recursive self-improvement looks 50:50 — ben_j_todd · 2026-09-29