The paper investigates when a language model actually learns task information versus when its readout (unembedding matrix) learns to express that information. Using a new parameter‑trajectory cross‑coding method, the authors find that hidden states can encode task data early in training, but the readout reorganizes only in a narrow window (0.4–1.1% of Pythia’s pretraining). By varying the readout learning rate, they show that faster readout learning reduces the gap between hidden‑state encoding and output expression without affecting final performance.
By Matteo He, William F. Shen, Alex Iacob, Andrej Jovanovic, Xinchi Qiu, Nicholas D. Lane
arXiv:2606. 01060v1 Announce Type: cross Abstract: Preference alignment has substantially improved the observable behavior of large language models, yet it remains unclear what alignment changes internally.
By Partha Pratim Saha, Samarth Raina, Mayur Parvatikar, Amit Dhanda, Vinija Jain, Aman Chadha, Amitava Das
arXiv:2610.19041v1 Announce Type: new
Abstract: Understanding the dynamics of language acquisition in Large Language Models (LLMs) is crucial for uncovering how statistical representations evolve int...
By Forough Shirin Abkenar, Haoshuo Fu, Liang Feng, Ning Yan, Masood S. Mortazavi
arXiv:2506.17871v4 Announce Type: replace-cross
Abstract: Despite their impressive capabilities, aligned large language models (LLMs) often generate outputs that lack diversity. What drives this cons...
By Chenghao Yang, Sida Li, Ari Holtzman
Large language models generate one token at a time, yet their responses show remarkably consistent length structure: step-by-step solutions converge in predictable token counts, retrievals stop after a few sentences, retractions extend responses by measurable amounts. We ask whether the model carries an internal estimate of how much response remains.
arXiv:2607. 05316v1 Announce Type: cross Abstract: Large language models generate one token at a time, yet their responses show remarkably consistent length structure: step-by-step solutions converge in predictable token counts, retrievals stop after a few sentences, retractions extend responses by measurable amounts.
By Mohamed Amine Merzouk, Dmitri Carpov, Mirko Bronzi, Damiano Fornasiere, Adam Oberman
arXiv:2607. 18305v1 Announce Type: cross Abstract: Some limits on what language models know are not gaps in data coverage but structural properties of learning from text.
By Priyansh Srivastava, Romit Chatterjee
arXiv:2606. 07559v1 Announce Type: cross Abstract: Fine-tuning a language model on contexts whose correct completion has a near-synonym competitor often fails silently.
By Vaibhav Prakash, Jayasri Dontabhaktuni
arXiv:2608.30315v1 Announce Type: new
Abstract: Token embeddings are the basic representational units that connect discrete tokens with continuous computation in language models. Although modern lang...
By Junjie Yao, Liangkai Hang, Zhi-Qin John Xu
arXiv:2510. 01163v2 Announce Type: replace Abstract: The factors driving the performance of in-context learning (ICL) in large language models (LLMs) remain poorly understood despite ICL's surprising effectiveness, enabling models to adapt to new tasks from only a handful of examples.
By Wa\"iss Azizian, Ali Hasan
The paper investigates how large language models (LLMs) share a common Fisher‑Rao geometry in their next‑token probability distributions, revealing that behaviour largely determines this geometry while activation geometry depends on coordinate choices. Across transformer, state‑space, and recurrent architectures, output geometries align more closely than activation geometries, and this shared structure facilitates semantic‑category transfer and improves agreement with human word choices as models scale and train. The study further demonstrates that geometry can guide minimum‑disturbance interventions, enabling reusable control that preserves behaviour better than Euclidean methods and enhances steering, editing, attribution, dictionary learning, and fine‑tuning.
By Dario Picozzi
arXiv:2603.18908v5 Announce Type: replace
Abstract: Independently trained language models often learn compatible late-stage representations, despite differences in training objectives, architectures,...
By Matt Gorbett, Suman Jana