The Natural Language Learning (NLL) reading group is a student-run group at UT Austin and part of the
UT Austin NLP and Computational Linguistics community.
We discuss recent papers in NLP, ML, and Computational Linguistics, and occasionally host invited talks.
All students, postdocs, and faculty interested in
language technologies are welcome to join!
The group is currently organized by Juan Diego Rodriguez.
January 30, . Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering. arXiv 2025 (pdf)
February 13, . FollowIR: Evaluating and Teaching Information Retrieval Models to Follow Instructions. NAACL 2025 (pdf)
February 27, . Signatures of human-like processing in Transformer forward passes. arXiv 2025 (pdf)
March 13, . Latent Collaboration in Multi-Agent Systems. arXiv 2025 (pdf)
March 27 - Matthew Finlayson (invited talk)
April 10, . Features as Rewards: Scalable Supervision for Open-Ended Tasks via Interpretability. arXiv 2026 (pdf)
April 24, . Small Reward Models via Backward Inference. arXiv 2026 (pdf) (Invited talk)