arXiv Computation and Language
Natural Language Autoencoders Identify Without Explaining
Read the original on arXiv Computation and Language →The Flow has not summarised this story yet — read it at arXiv Computation and Language.
Related stories
arXiv Computation and Language
Oct 23
Verbalizing Activation Differences with Natural Language Autoencoders
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.
Hugging Face Trending Papers
Oct 22
Verbalizing Activation Differences with Natural Language Autoencoders
The paper investigates natural language autoencoders (NLAs) that convert large language model activations into readable text. 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 multiple models show positive results: NLAs preserve compositionality, transfer to lightweight fine‑tunes, recover hidden objectives, and outperform a black‑box baseline on diffing tasks, though they lag behind state‑of‑the‑art methods for misalignment detection.
arXiv Machine Learning
Oct 21
On the Monitorability of Natural Language Autoencoders: Identification vs Discrimination
arXiv:2610.18727v2 Announce Type: replace Abstract: Natural Language Autoencoders (NLAs) are proposed as oversight tools for language models: an Activation Verbaliser (AV) writes an English explanati...
arXiv AI
Oct 26
Do Language Models Trust Their Own Justifications? A Study on Functional Consistency
The paper investigates whether the tokens highlighted by large language models (LLMs) as key to their predictions are behaviorally consistent when those tokens are manipulated. Using sufficiency, comprehensiveness, and counterfactuality metrics, the authors evaluate closed‑source GPT‑4o‑mini and open‑source models Gemma3, Granite8B, and DeepSeek on IMDB and Steam review datasets. Results show that GPT‑4o‑mini consistently follows expected patterns, Gemma3 and Granite8B maintain sufficiency but falter under stricter interventions, and DeepSeek variants exhibit structural inconsistencies, indicating that explanation reliability varies across model families and sizes.
arXiv AI
Jul 23
Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations
arXiv:2607. 20379v1 Announce Type: new Abstract: Natural-language autoencoders score explanations of hidden activations by reconstruction: an explanation is deemed faithful if the activation can be regenerated from it.
arXiv Computation and Language
Aug 25
Bridging Linguistic Structure and Mechanistic Interpretability for Conceptual Interpretation in Language Models
arXiv:2408.11827v2 Announce Type: replace Abstract: Understanding how language models compose meaning from linguistic input remains a central problem in interpretability research. Mechanistic studies...