Similar Items: Position: Mechanistic Interpretability Must Disclose Identification Assumptions for Causal Claims
- Superposition Is Not Necessary: A Mechanistic Interpretability Analysis of Transformer Representations for Time Series Forecasting
- Where's the Plan? Locating Latent Planning in Language Models with Lightweight Mechanistic Interventions
- Interpreting Reinforcement Learning Agents with Susceptibilities
- On the Wasserstein Gradient Flow Interpretation of Drifting Models
- Semiparametric Efficient Test for Interpretable Distributional Treatment Effects
- Bayesian Sensitivity of Causal Inference Estimators under Evidence-Based Priors