Elamur Research⌕
☰
home / authors / Andrew Lee

Andrew Lee

Andrew Lee has 3 papers in our library, in 2023–2026, most of them on Interpretability, Bias, fairness & safety; often with Martin Wattenberg, Aaron Mueller, Ekdeep Singh Lubana.

3 papers in our library · 2023–2026

OpenAlex · ORCID · h-index 4 · 17 works · 879 citations OpenAlex, 2026-10-14

📄 3 works in our library · 🔗 Cited 2,370 times in our library · ⭐ Paper Stars ranking: #6689

The profile is the library's match of the name; two people with one name are not always told apart.

Elsewhere: Google Scholar · Semantic Scholar · dblp · ACL Anthology

 Affiliations

🎓University of California, DavisDavis · United States2025 · 4 papersEmployment🎓Harvard UniversityCambridge · United States2023–2025 · 3 papers🎓University of Maryland, Baltimore CountyBaltimore · United States2016–2024 · 2 papers

From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 2 more institutions on single bylines are not listed.

 Fields, topics and concepts

Of its works in our library, with the number of works on each.

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Artificial Intelligence · 3 worksComputer Science · 3 worksSafety Research · 1 workSocial Sciences · 1 workArts and Humanities · 1 workLanguage and Linguistics · 1 work

🧭 Topics 2Research subjects and languages the work is about.Interpretability · 3 worksBias, fairness & safety · 1 work

 Works by year

2020s

20261 paper20241 paper20231 paper

 Papers

The most cited first (OpenAlex and our library).

  1. Emergent Linear Representations in World Models of Self-Supervised Sequence Models (2023 · 116 citations)
    This paper asks whether the board-state "world model" learned by OthelloGPT is encoded linearly. Li et al. (2023a) found that only non-linear probes could read it. The authors argue that the model does not track absolute tile colours (BLACK/WHITE/EMPTY). Instead, it tracks colours relative to whoever is moving at the…
  2. A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity (2024 · 64 citations)
    Alignment methods based on preference tuning are now routinely used to make language models safer. However, it is unclear what actually changes inside a model when it becomes "aligned", and so jailbreaks remain poorly explained. This paper uses toxicity reduction with direct preference optimization (DPO) in…
  3. From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts? (2026)
    Interpretability research usually judges featurizers such as sparse autoencoders (SAEs) and probes one concept at a time, implicitly assuming that concepts are independent.

 Co-authors here

6 co-authors with a page, the most papers together first.

Venues

Where their works here were published, the most works first, with the editions of those that have a page.

 Cited by

Papers of our library cite their works 2,370 times (of the 149,439 papers whose bibliographies we read); 594 citing papers with a page here →.

Updated

This page in Markdown → https://md.research.elamur.ai/free/en/authors/andrew-lee/