Research map / Architectures and efficiency / RNNs and normalization
Influence Dynamics and Stagewise Data Attribution
Jin Hwa Lee, Matthew J. A. Smith, Maxwell Adam, Jesse Hoogland
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| measure influence at checkpoints during training (stagewise BIF) rather than a single final-model influence | deep linear network on hierarchical dataset · Toy hierarchical semantic dataset | influence sign and magnitude over time | Improved |
| compare BIF with leave-one-out retraining and analytical influence | BIF in toy model · Toy hierarchical semantic dataset | agreement of influence trajectories | Improved |
| ablate a sample only during a short window at the BIF peak (ablation) | deep linear network (dog sample) · Toy hierarchical semantic dataset | integrated loss difference | Improved |
| upweight induction patterns when the induction circuit begins to form versus earlier (ablation) | small language model trained from scratch · Induction head formation | induction head formation speed | Improved |
| use high localization gamma and low step size in SGLD (ablation) | BIF estimator · Toy model BIF-LOO correlation | Pearson correlation with LOO | Improved |
| use RMSProp-SGLD instead of vanilla SGLD (ablation) | BIF/LLC estimator · Deep linear network LLC | match to ground-truth LLC | Improved |
Closest papers by meaning
- Capturing the Temporal Dependence of Training Data Influence (ICLR 2025 · similarity 0.85)
- Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space (NeurIPS 2024 · similarity 0.85)
- The Mirrored Influence Hypothesis: Efficient Data Influence Estimation by Harnessing Forward Passes (CVPR 2024 · similarity 0.84)
- Sketching the Readout of Large Language Models for Scalable Data Attribution and Valuation (NeurIPS 2026 · similarity 0.84)
- Influence Functions for Scalable Data Attribution in Diffusion Models (ICLR 2025 · similarity 0.84)
- Language Model Behavioral Phases are Consistent Across Architecture, Training Data, and Scale (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
