Tag: ai-safety
-
Neural Circuit Cartography & AI Safety Architectures
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in Neural Circuit Cartography & AI Safety Architectures.
-
AI Sleeper Agents & Deceptive Alignment Verification
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in AI Sleeper Agents & Deceptive Alignment Verification.
-
ROME Algorithm: Rank-One Model Editing in Neural Memory
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in ROME Algorithm: Rank-One Model Editing in Neural Memory.
-
Activation Patching & Causal Mediation in Deep Models
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in Activation Patching & Causal Mediation in Deep Models.
-
Sparse Autoencoders for Mechanistic Feature Extraction
Disentangling polysemantic neuron superposition into interpretable, atomic feature directions.
-
The Superposition Hypothesis & Polysemantic Neurons
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in The Superposition Hypothesis & Polysemantic Neurons.
-
QK vs OV Circuits in Transformer Multi-Head Attention
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in QK vs OV Circuits in Transformer Multi-Head Attention.
-
Mechanistic Induction Heads & In-Context Pattern Copying
How two-layer attention subcircuits implement in-context pattern replication and prefix matching.
-
The Residual Stream in Scaled Transformer Architectures
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in The Residual Stream in Scaled Transformer Architectures.
-
Foundations of Mechanistic Interpretability in AI
Physical mechanisms, mathematical boundary conditions, and architectural trade-offs in Foundations of Mechanistic Interpretability in AI.