Aaquib Syed
Aaquib Syed has 2 papers in our library, in 2023–2024, most of them on Language models & pretraining, Interpretability, Bias, fairness & safety; often with Arthur Conmy, Can Rager, Daniel Paleka.
OpenAlex · h-index 3 · 3 works · 84 citations OpenAlex, 2026-10-14
📄 2 works in our library · 🔗 Cited 256 times in our library · ⭐ Paper Stars ranking: #5739
Fields, topics and concepts
📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 2 worksArtificial Intelligence · 2 worksArts and Humanities · 1 workLanguage and Linguistics · 1 workComputer Vision and Pattern Recognition · 1 workSoftware · 1 work
🧭 Topics 3Research subjects and languages the work is about.Interpretability · 1 workBias, fairness & safety · 1 workLanguage models & pretraining · 1 work
Works by year
2020s
Papers
- Refusal in Language Models Is Mediated by a Single Direction (2024 · 196 citations)
Chat LLMs are safety-tuned to decline harmful requests, but how this behavior works inside the model is poorly understood. - Attribution Patching Outperforms Automated Circuit Discovery (2023 · 81 citations)
Mechanistic interpretability tries to find circuits, which are small subgraphs of a network that account for a specific behavior. Automated tools such as ACDC rely on activation patching. That method needs an extra forward pass for each edge it tests, so it scales poorly to large models.
Co-authors here
Cited by
Updated
This page in Markdown → https://md.research.elamur.ai/free/en/authors/aaquib-syed/