🧩 Philosophy 4d ago · RGRGRG

Computation in Superposition: Two Handcrafted Models

Less Wrong
View Channel →
Computation in Superposition: Two Handcrafted Models
Source ↗ 👁 0 💬 0
Many interpretability researchers (ourselves included) believe that neural networks store knowledge in superposition—that is, networks encode more facts than they have individual components. A natural extension of this idea is that networks also perform computation on knowledge that lives in superposition. Despite the centrality of this concept, there are few concrete examples of what computation in superposition actually looks like in practice.In this post, we study a toy memorization task wher

Comments (0)

Sign in to join the discussion

More Like This

Online Philosophy Resources Weekly Update
Daily Nous · 2h ago
Protocol for a First Date
3:AM Magazine · 3h ago
📰
Friedrich Albert Lange
Stanford Encyclopedia of Philosophy · 6h ago
📰
Conflict 2.0: Leaving behind shame/fault, right/wrong
LessWrong · 8h ago
Enneagram Epicycles
LessWrong · 8h ago
📰
Auto-review of agent actions without synchronous human oversight
LessWrong · 9h ago