Methods · Learning Dynamics and Mechanistic Interpretability

Trajectory-Based Attribution

Survey sections: §10.2.1 Trajectory-Based Attribution

Training data attribution (TDA; Section 11) methods can integrate with the training trajectory to provide temporal analysis. For example, TracIn (Section 11.2.5) estimates a training sample’s influence on a testing sample’s prediction as the sum of their gradients’ dot products evaluated at checkpoints along the run (Garima et al., 2020).
Survey §10.2.1 Trajectory-Based Attribution

Intended applications

Each checkpoint contributes separately to the influence estimate. By examining each checkpoint’s contribution, researchers can identify when a training example had the greatest influence on a downstream prediction.

Related methods

Compare side by side →

Key papers

Cited in §10.2.1.

Garima, Frederick Liu, Satyen Kale, and Mukund Sundararajan. Estimating training data influence by tracing gradient descent. In Proceedings of the 34th International Conference on Neural Information Processing Systems, NIPS ’20, Red Hook, NY, USA, 2020. Curran Associates Inc. ISBN 9781713829546.

Other methods in Learning Dynamics and Mechanistic Interpretability

Per-Token Learning Curves · Refined Local Learning Coefficients (Localising Change to Components) · Forked-Run Experiments and Linear Mode Connectivity · Loss-Curve Derivatives · Capability-Specific Metrics · Per-Example Decomposition (POLCA) · Latent State Models · Progress Measures · Local Learning Coefficient (LLC) · Forecasting from Smaller and Partial Runs · Forecasting Safety-Relevant Behaviours · Internal Precursors as Early Warning · Pretraining Data Filtering (Deep Ignorance) · Gradient Routing · Monitoring and Steering During Fine-Tuning (Persona Vectors) · Train-Time Intervention Loop