The paper distinguishes between active facts, which help a language model generalize beyond mere recall, and inert facts, which do not. By relabeling answers to a known relation (e.g., Paris → Europe becomes Paris → crab) and fine‑tuning on a subset of subjects, the authors test whether the model can transfer the new label to unseen subjects sharing the same original answer. Across eleven models and 70 relations, the linear readability of the original relation in pretrained representations strongly predicts this generalization, with a Pearson correlation of 0.87–0.92, even after controlling for various factors.
arXiv:2409. 02228v2 Announce Type: replace Abstract: When language models (LMs) are trained to forget (or "unlearn'') a skill, how precisely does their behavior change?
By Eric Zhang, Leshem Choshen, Jacob Andreas
arXiv:2606.00284v2 Announce Type: replace
Abstract: While continual pretraining (CPT) is a practical way to extend large language models to new languages, na\"ive finetuning often erodes existing cap...
By Sanchit Ahuja, Terra Blevins
arXiv:2606. 11198v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) systems inject external knowledge to improve LLM outputs, yet the format of injected content -- distinct from its semantic relevance -- can independently distort the model's attention distribution.
By Yuqi Zhang, Di Zhang
arXiv:2610.27724v1 Announce Type: new
Abstract: Dense linear probes assess task readability from language model (LM) representations, but do not directly reveal how many features suffice for predicti...
By Kaiyang Wan, Yuxia Wang
arXiv:2606. 00467v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used for zero-shot annotation and LLM-as-a-judge tasks, yet their reliability hinges on how model-internalized priors interact with user-provided instructions.
By Etienne Casanova, Rafal Kocielnik, R. Michael Alvarez
LMEnt is a toolkit designed to study how language models acquire and represent world knowledge. It includes a richly annotated pretraining corpus, an improved entity-based retrieval method, and twelve pretrained models with numerous checkpoints. The suite enables controlled experiments linking entity mentions in training data to downstream knowledge performance, revealing how co-occurrence patterns influence learning and editing difficulty.
By Daniela Gottesman, Alon Gilae-Dotan, Ido Cohen, Yoav Gur-Arieh, Marius Mosbach, Ori Yoran, Mor Geva
Activation verbalizers translate the internal representations of large language models into human-readable text. The recently proposed natural language autoencoders (NLAs) learn this by reconstructing...
arXiv:2610.18674v1 Announce Type: cross
Abstract: Large language models rely on prompt context to perform in-context learning, follow formats, and adapt to local evidence. This ability, however, also...
By Jinghan Zheng, Jiajun Shi
arXiv:2609.23065v1 Announce Type: cross
Abstract: Chain-of-thought (CoT) can sound plausible yet be unfaithful to the model's underlying reasoning. Most prior work probes CoT faithfulness through inp...
By Qianli Wang, Yilong Wang, Dennis Wei, Jingyi Sun, Simon Ostermann, Pepa Atanasova, Nils Feldhus
The paper investigates natural language autoencoders (NLAs) as tools for verbalizing activation differences in large language models. It examines whether NLAs capture compositional structure, generalize to fine‑tuned models, recover fine‑tuning objectives, and can be specialized for model diffing. Experiments across four models show that NLAs largely succeed in these tasks, though they do not yet match state‑of‑the‑art diffing methods on emergent misalignment detection.
By Andrea Blasi N\'u\~nez, Gianluca Barmina, Federico Torrielli, Thiusius Rajeeth Savarimuthu, Peter Schneider-Kamp, Lukas Galke Poech
arXiv:2609.39702v1 Announce Type: new
Abstract: Adapting a language model to a specialized corpus means choosing which instruction-tuning tasks to train on under a fixed budget, and testing one choic...
By Nima H. Siboni, Vahid Rostami