# First is Better Than Last for Language Data Influence

arXiv: https://arxiv.org/abs/2202.11844 · HTML: https://mlsanity.com/paper/2202.11844/

## Key facts

- Cited by 34 papers in the map (data through 14 Oct 2026).
- Builds on 36 of its 59 references that are in the map.
- First posted on arXiv 24 Feb 2022.

## Builds on (most cited first, 30 of 36)

- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://mlsanity.com/paper/2006.03654/) (2020)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1704.02685/) (2017)
- [Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)](https://mlsanity.com/paper/1711.11279/) (2017)
- [Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPR](https://mlsanity.com/paper/1711.00399/) (2017)
- [RISE: Randomized Input Sampling for Explanation of Black-box Models](https://mlsanity.com/paper/1806.07421/) (2018)
- [This Looks Like That: Deep Learning for Interpretable Image Recognition](https://mlsanity.com/paper/1806.10574/) (2018)
- [Estimating Training Data Influence by Tracing Gradient Descent](https://mlsanity.com/paper/2002.08484/) (2020)
- [Counterfactual Visual Explanations](https://mlsanity.com/paper/1904.07451/) (2019)
- [On the (In)fidelity and Sensitivity for Explanations](https://mlsanity.com/paper/1901.09392/) (2019)
- [Towards Efficient Data Valuation Based on the Shapley Value](https://mlsanity.com/paper/1902.10275/) (2019)
- [Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negatives](https://mlsanity.com/paper/1802.07623/) (2018)
- [Representer Point Selection for Explaining Deep Neural Networks](https://mlsanity.com/paper/1811.09720/) (2018)
- [On the Accuracy of Influence Functions for Measuring Group Effects](https://mlsanity.com/paper/1905.13289/) (2019)
- [Explaining Black Box Predictions and Unveiling Data Artifacts through Influence Functions](https://mlsanity.com/paper/2005.06676/) (2020)
- [Guided Integrated Gradients: An Adaptive Path Method for Removing Noise](https://mlsanity.com/paper/2106.09788/) (2021)
- [Grounding Visual Explanations](https://mlsanity.com/paper/1807.09685/) (2018)
- [AllenNLP Interpret: A Framework for Explaining Predictions of NLP Models](https://mlsanity.com/paper/1909.09251/) (2019)
- [Falling Rule Lists](https://mlsanity.com/paper/1411.5899/) (2014)
- [Towards better understanding of gradient-based attribution methods for Deep Neural Networks](https://mlsanity.com/paper/1711.06104/) (2017)
- [Interpreting Black Box Predictions using Fisher Kernels](https://mlsanity.com/paper/1810.10118/) (2018)
- [Revisiting Methods for Finding Influential Examples](https://mlsanity.com/paper/2111.04683/) (2021)

## Built on by (most cited first, 30 of 34)

- [Training Data Influence Analysis and Estimation: A Survey](https://mlsanity.com/paper/2212.04612/) (2022)
- [What is Your Data Worth to GPT? LLM-Scale Data Valuation with Influence Functions](https://mlsanity.com/paper/2405.13954/) (2024)
- [DoGE: Domain Reweighting with Generalization Estimation](https://mlsanity.com/paper/2310.15393/) (2023)
- [Massive Editing for Large Language Models via Meta Learning](https://mlsanity.com/paper/2311.04661/) (2023)
- [When Search Engine Services meet Large Language Models: Visions and Challenges](https://mlsanity.com/paper/2407.00128/) (2024)
- [Scalable Influence and Fact Tracing for Large Language Model Pretraining](https://mlsanity.com/paper/2410.17413/) (2024)
- [Training Data Attribution via Approximate Unrolled Differentiation](https://mlsanity.com/paper/2405.12186/) (2024)
- [Improving Influence-based Instruction Tuning Data Selection for Balanced Learning of Diverse Capabilities](https://mlsanity.com/paper/2501.12147/) (2025)
- [Make Every Example Count: On the Stability and Utility of Self-Influence for Learning from Noisy NLP Datasets](https://mlsanity.com/paper/2302.13959/) (2023)
- [Helpful or Harmful Data? Fine-tuning-free Shapley Attribution for Explaining Language Model Predictions](https://mlsanity.com/paper/2406.04606/) (2024)
- [Unifying Corroborative and Contributive Attributions in Large Language Models](https://mlsanity.com/paper/2311.12233/) (2023)
- [Efficient Sketches for Training Data Attribution and Studying the Loss Landscape](https://mlsanity.com/paper/2402.03994/) (2024)
- [Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety](https://mlsanity.com/paper/2506.05451/) (2025)
- [Towards Explainable Artificial Intelligence (XAI): A Data Mining Perspective](https://mlsanity.com/paper/2401.04374/) (2024)
- [In2Core: Leveraging Influence Functions for Coreset Selection in Instruction Finetuning of Large Language Models](https://mlsanity.com/paper/2408.03560/) (2024)
- [LearnAlign: Data Selection for LLM Reinforcement Learning with Improved Gradient Alignment](https://mlsanity.com/paper/2506.11480/) (2025)
- [MergeMix: Optimizing Mid-Training Data Mixtures via Learnable Model Merging](https://mlsanity.com/paper/2601.17858/) (2026)
- [Skill-Targeted Adaptive Training](https://mlsanity.com/paper/2510.10023/) (2025)
- [First is Not Really Better Than Last: Evaluating Layer Choice and Aggregation Strategies in Language Model Data Influence Estimation](https://mlsanity.com/paper/2511.04715/) (2025)
- [AutoMixer: Checkpoint Artifacts as Automatic Data Mixers](https://mlsanity.com/paper/2506.21910/) (2025)
- [Influential Training Data Retrieval for Explaining Verbalized Confidence of LLMs](https://mlsanity.com/paper/2601.10645/) (2026)
- [Learning to Weight Parameters for Training Data Attribution](https://mlsanity.com/paper/2506.05647/) (2025)
- [Fast-DataShapley: Neural Modeling for Training Data Valuation](https://mlsanity.com/paper/2506.05281/) (2025)
- [Influence-driven Curriculum Learning for Pre-training on Limited Data](https://mlsanity.com/paper/2508.15475/) (2025)
- [SoK: Memorisation in machine learning](https://mlsanity.com/paper/2311.03075/) (2023)
- [Golden Layers and Where to Find Them: Improved Knowledge Editing for Large Language Models Via Layer Gradient Analysis](https://mlsanity.com/paper/2602.20207/) (2026)
- [DelTA: Discriminative Token Credit Assignment for Reinforcement Learning from Verifiable Rewards](https://mlsanity.com/paper/2605.21467/) (2026)
- [STRIDE: Training Data Attribution via Sparse Recovery from Subset Perturbations](https://mlsanity.com/paper/2606.05165/) (2026)
- [Global-to-Local Support Spectrums for Language Model Explainability](https://mlsanity.com/paper/2408.05976/) (2024)
- [Uncovering Scaling Laws for Large Language Models via Inverse Problems](https://mlsanity.com/paper/2509.07909/) (2025)

## Abstract (arXiv)

The ability to identify influential training examples enables us to debug training data and explain model behavior. Existing techniques to do so are based on the flow of training data influence through the model parameters. For large models in NLP applications, it is often computationally infeasible to study this flow through all model parameters, therefore techniques usually pick the last layer of weights. However, we observe that since the activation connected to the last layer of weights contains "shared logic", the data influenced calculated via the last layer weights prone to a ``cancellation effect'', where the data influence of different examples have large magnitude that contradicts each other. The cancellation effect lowers the discriminative power of the influence score, and deleting influential examples according to this measure often does not change the model's behavior by much. To mitigate this, we propose a technique called TracIn-WE that modifies a method called TracIn to operate on the word embedding layer instead of the last layer, where the cancellation effect is less severe. One potential concern is that influence based on the word embedding layer may not encode sufficient high level information. However, we find that gradients (unlike embeddings) do not suffer from this, possibly because they chain through higher layers. We show that TracIn-WE significantly outperforms other data influence methods applied on the last layer significantly on the case deletion evaluation on three language classification tasks for different models. In addition, TracIn-WE can produce scores not just at the level of the overall training input, but also at the level of words within the training input, a further aid in debugging.