This paper presents Markovian Circuit Tracing (MCT) as a benchmark and pipeline to extract and test state-transition structures in transformer activations using synthetic HMM tasks, demonstrating that state patching improves counterfactual predictions.
Towards monosemanticity: Decomposing language models with dictionary learning
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2representative citing papers
SwiftGS predicts satellite 3D surfaces and renderings zero-shot via meta-learned Gaussian-SDF hybrid, reporting 1.22 m DSM MAE on DFC2019 at 2.5 min per scene.
citing papers explorer
-
Markovian Circuit Tracing for Transformer State Dynamic
This paper presents Markovian Circuit Tracing (MCT) as a benchmark and pipeline to extract and test state-transition structures in transformer activations using synthetic HMM tasks, demonstrating that state patching improves counterfactual predictions.
-
SwiftGS: Episodic Priors for Immediate Satellite Surface Recovery
SwiftGS predicts satellite 3D surfaces and renderings zero-shot via meta-learned Gaussian-SDF hybrid, reporting 1.22 m DSM MAE on DFC2019 at 2.5 min per scene.