# Start Making Sense(s): A Developmental Probe of Attention Specialization Using Lexical Ambiguity

arXiv: https://arxiv.org/abs/2511.21974 · HTML: https://mlsanity.com/paper/2511.21974/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 36 of its 50 references that are in the map.
- First posted on arXiv 26 Nov 2025.

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

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling](https://mlsanity.com/paper/2304.01373/) (2023)
- [Enriching Word Vectors with Subword Information](https://mlsanity.com/paper/1607.04606/) (2016)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [What Does BERT Look At? An Analysis of BERT's Attention](https://mlsanity.com/paper/1906.04341/) (2019)
- [In-context Learning and Induction Heads](https://mlsanity.com/paper/2209.11895/) (2022)
- [Are Sixteen Heads Really Better than One?](https://mlsanity.com/paper/1905.10650/) (2019)
- [Towards Automated Circuit Discovery for Mechanistic Interpretability](https://mlsanity.com/paper/2304.14997/) (2023)
- [Revealing the Dark Secrets of BERT](https://mlsanity.com/paper/1908.08593/) (2019)
- [Stress Test Evaluation for Natural Language Inference](https://mlsanity.com/paper/1806.00692/) (2018)
- [Open Problems in Mechanistic Interpretability](https://mlsanity.com/paper/2501.16496/) (2025)
- [Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Models](https://mlsanity.com/paper/2305.14763/) (2023)
- [Sudden Drops in the Loss: Syntax Acquisition, Phase Transitions, and Simplicity Bias in MLMs](https://mlsanity.com/paper/2309.07311/) (2023)
- [Inducing Causal Structure for Interpretable Neural Networks](https://mlsanity.com/paper/2112.00826/) (2021)
- [SemEval-2020 Task 1: Unsupervised Lexical Semantic Change Detection](https://mlsanity.com/paper/2007.11464/) (2020)
- [What Matters in Transformers? Not All Attention is Needed](https://mlsanity.com/paper/2406.15786/) (2024)
- [LLM Circuit Analyses Are Consistent Across Training and Scale](https://mlsanity.com/paper/2407.10827/) (2024)
- [Talking Heads: Understanding Inter-layer Communication in Transformer Language Models](https://mlsanity.com/paper/2406.09519/) (2024)
- [Let's Play Mono-Poly: BERT Can Reveal Words' Polysemy Level and Partitionability into Senses](https://mlsanity.com/paper/2104.14694/) (2021)
- [PolyPythias: Stability and Outliers across Fifty Language Model Pre-Training Runs](https://mlsanity.com/paper/2503.09543/) (2025)
- [The Same But Different: Structural Similarities and Differences in Multilingual Language Modeling](https://mlsanity.com/paper/2410.09223/) (2024)
- [Contextualized Word Embeddings Encode Aspects of Human-Like Word Sense Knowledge](https://mlsanity.com/paper/2010.13057/) (2020)
- [Latent State Models of Training Dynamics](https://mlsanity.com/paper/2308.09543/) (2023)
- [Does Time Have Its Place? Temporal Heads: Where Language Models Recall Time-specific Information](https://mlsanity.com/paper/2502.14258/) (2025)
- [Position-aware Automatic Circuit Discovery](https://mlsanity.com/paper/2502.04577/) (2025)
- [RAW-C: Relatedness of Ambiguous Words--in Context (A New Lexical Resource for English)](https://mlsanity.com/paper/2105.13266/) (2021)
- [Exploring the Representation of Word Meanings in Context: A Case Study on Homonymy and Synonymy](https://mlsanity.com/paper/2106.13553/) (2021)
- [The DURel Annotation Tool: Human and Computational Measurement of Semantic Proximity, Sense Clusters and Semantic Change](https://mlsanity.com/paper/2311.12664/) (2023)

## Built on by (most cited first, 3 of 3)

- [Position: Don't Just "Fix it in Post": A Science of AI Must Study Training Dynamics](https://mlsanity.com/paper/2606.06533/) (2026)
- [Capacity Constraints and the Multilingual Penalty for Lexical Disambiguation](https://mlsanity.com/paper/2602.05035/) (2026)
- [Developmental Trajectories of Situation Modeling and Mentalizing in Transformer Language Models](https://mlsanity.com/paper/2606.28524/) (2026)

## Abstract (arXiv)

Despite an in-principle understanding of self-attention matrix operations in Transformer language models (LMs), it remains unclear precisely how these operations map onto interpretable computations or functions--and how or when individual attention heads develop specialized attention patterns. Here, we present a pipeline to systematically probe attention mechanisms, and we illustrate its value by leveraging lexical ambiguity--where a single word has multiple meanings--to isolate attention mechanisms that contribute to word sense disambiguation. We take a "developmental" approach: first, using publicly available Pythia LM checkpoints, we identify inflection points in disambiguation performance for each LM in the suite; in 14M and 410M, we identify heads whose attention to disambiguating words covaries with overall disambiguation performance across development. We then stress-test the robustness of these heads to stimulus perturbations: in 14M, we find limited robustness, but in 410M, we identify multiple heads with surprisingly generalizable behavior. Then, in a causal analysis, we find that ablating the target heads demonstrably impairs disambiguation performance, particularly in 14M. We additionally reproduce developmental analyses of 14M across all of its random seeds. Together, these results suggest: that disambiguation benefits from a constellation of mechanisms, some of which (especially in 14M) are highly sensitive to the position and part-of-speech of the disambiguating cue; and that larger models (410M) may contain heads with more robust disambiguation behavior. They also join a growing body of work that highlights the value of adopting a developmental perspective when probing LM mechanisms.