# Dual Process Learning: Controlling Use of In-Context vs. In-Weights Strategies with Weight Forgetting

arXiv: https://arxiv.org/abs/2406.00053 · HTML: https://mlsanity.com/paper/2406.00053/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 54 references that are in the map.
- First posted on arXiv 28 May 2024.

## Builds on (most cited first, 22 of 22)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [BERT Rediscovers the Classical NLP Pipeline](https://mlsanity.com/paper/1905.05950/) (2019)
- [Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books](https://mlsanity.com/paper/1506.06724/) (2015)
- [What Can Transformers Learn In-Context? A Case Study of Simple Function Classes](https://mlsanity.com/paper/2208.01066/) (2022)
- [Measuring Catastrophic Forgetting in Neural Networks](https://mlsanity.com/paper/1708.02072/) (2017)
- [Toward Transparent AI: A Survey on Interpreting the Inner Structures of Deep Neural Networks](https://mlsanity.com/paper/2207.13243/) (2022)
- [The mechanistic basis of data dependence and abrupt learning in an in-context classification task](https://mlsanity.com/paper/2312.03002/) (2023)
- [The MultiBERTs: BERT Reproductions for Robustness Analysis](https://mlsanity.com/paper/2106.16163/) (2021)
- [The Transient Nature of Emergent In-Context Learning in Transformers](https://mlsanity.com/paper/2311.08360/) (2023)
- [Conditional probing: measuring usable information beyond a baseline](https://mlsanity.com/paper/2109.09234/) (2021)
- [Transformers generalize differently from information stored in context vs in weights](https://mlsanity.com/paper/2210.05675/) (2022)
- [Fishing for Magikarp: Automatically Detecting Under-trained Tokens in Large Language Models](https://mlsanity.com/paper/2405.05417/) (2024)
- [Neural Collapse in the Intermediate Hidden Layers of Classification Neural Networks](https://mlsanity.com/paper/2308.02760/) (2023)
- [The Impact of Reinitialization on Generalization in Convolutional Neural Networks](https://mlsanity.com/paper/2109.00267/) (2021)
- [Knowledge Evolution in Neural Networks](https://mlsanity.com/paper/2103.05152/) (2021)
- [Amnesic Probing: Behavioral Explanation with Amnesic Counterfactuals](https://mlsanity.com/paper/2006.00995/) (2020)
- [Syntax Representation in Word Embeddings and Neural Networks -- A Survey](https://mlsanity.com/paper/2010.01063/) (2020)
- [A Neural Network Model of Continual Learning with Cognitive Control](https://mlsanity.com/paper/2202.04773/) (2022)

## Built on by (most cited first, 15 of 15)

- [Training Dynamics of In-Context Learning in Linear Attention](https://mlsanity.com/paper/2501.16265/) (2025)
- [Emergent Abilities in Large Language Models: A Survey](https://mlsanity.com/paper/2503.05788/) (2025)
- [Pixels Versus Priors: Controlling Knowledge Priors in Vision-Language Models through Visual Counterfacts](https://mlsanity.com/paper/2505.17127/) (2025)
- [Strategy Coopetition Explains the Emergence and Transience of In-Context Learning](https://mlsanity.com/paper/2503.05631/) (2025)
- [Investigating the Pre-Training Dynamics of In-Context Learning: Task Recognition vs. Task Learning](https://mlsanity.com/paper/2406.14022/) (2024)
- [Rapid Word Learning Through Meta In-Context Learning](https://mlsanity.com/paper/2502.14791/) (2025)
- [ICL CIPHERS: Quantifying "Learning" in In-Context Learning via Substitution Ciphers](https://mlsanity.com/paper/2504.19395/) (2025)
- [Softmax ≥ Linear: Transformers may learn to classify in-context by kernel gradient descent](https://mlsanity.com/paper/2510.10425/) (2025)
- [In-context Learning in Presence of Spurious Correlations](https://mlsanity.com/paper/2410.03140/) (2024)
- [Evaluating Steering Techniques using Human Similarity Judgments](https://mlsanity.com/paper/2505.19333/) (2025)
- [Large language models reorganize representational geometry during in-context learning](https://mlsanity.com/paper/2605.28854/) (2026)
- [To See the Unseen: on the Generalization Ability of Transformers in Symbolic Reasoning](https://mlsanity.com/paper/2604.21632/) (2026)
- [Information Abundance Paradox: Long-Context Training Undermines Parametric Knowledge](https://mlsanity.com/paper/2608.12218/) (2026)
- [Decomposing Prediction Mechanisms for In-Context Recall](https://mlsanity.com/paper/2507.01414/) (2025)
- [Reconciling In-Context and In-Weight Learning via Dual Representation Space Encoding](https://mlsanity.com/paper/2603.13459/) (2026)

## Abstract (arXiv)

Language models have the ability to perform in-context learning (ICL), allowing them to flexibly adapt their behavior based on context. This contrasts with in-weights learning (IWL), where memorized information is encoded in model parameters after iterated observations of data. An ideal model should be able to flexibly deploy both of these abilities. Despite their apparent ability to learn in-context, language models are known to struggle when faced with unseen or rarely seen tokens (Land & Bartolo, 2024). Hence, we study structural in-context learning, which we define as the ability of a model to execute in-context learning on arbitrary novel tokens -- so called because the model must generalize on the basis of e.g. sentence structure or task structure, rather than content encoded in token embeddings. We study structural in-context algorithms on both synthetic and naturalistic tasks using toy models, masked language models, and autoregressive language models. We find that structural ICL appears before quickly disappearing early in LM pretraining. While it has been shown that ICL can diminish during training (Singh et al., 2023), we find that prior work does not account for structural ICL. Building on Chen et al. (2024) 's active forgetting method, we introduce pretraining and finetuning methods that can modulate the preference for structural ICL and IWL. Importantly, this allows us to induce a textitdual process strategy where in-context and in-weights solutions coexist within a single model.