arXiv:2609. 12591v1 Announce Type: new Abstract: Foundation models are increasingly adapted through fine-tuning, model editing, and alignment procedures while retaining previously acquired capabilities.
arXiv:2607. 20596v1 Announce Type: new Abstract: Sparse autoencoder (SAE) features are used to interpret and steer large language models, yet whether a feature's causal role is stable across SAE families remains untested.
arXiv:2605. 28149v2 Announce Type: replace Abstract: Sparse Autoencoders (SAEs) extract interpretable features from Large Language Model activations, but standard variants enforce non-negative latents, so a bidirectional semantic axis (e.
arXiv:2606. 08365v1 Announce Type: cross Abstract: Sparse autoencoder (SAE) features are increasingly used to steer language models, but feature steering is rarely clean: the same intervention can behave inconsistently across contexts and perturb unrelated features.
arXiv:2607. 12166v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) are the standard for decomposing superposed neural representations into interpretable features, and evaluation relies predominantly on correlational recovery metrics -- cosine similarity between ground-truth directions and decoder atoms.
The paper introduces a new framework for evaluating sparse autoencoders (SAEs) by requiring their features to form an atomic interpretable unit basis, a coordinate system for semantically meaningful explanation. It formalizes this objective and organizes existing and new metrics, such as DAME, into a comprehensive benchmarking framework. Experiments across many SAE architectures reveal that some models appear monosemantic under current metrics yet are functionally polysemantic, with dominant features being less interpretable and more redundant.
arXiv:2606. 18383v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) are increasingly used to extract interpretable features from language models (LMs), yet a central question remains: when can an SAE-based explanation be treated as a faithful view of an underlying frozen LM We study this through a post-hoc generalization framework that certifies the LM via a sparse proxy, obtained by replacing a native hidden activation with its pretrained SAE reconstruction.
arXiv:2604. 26866v2 Announce Type: replace-cross Abstract: Large language models (LLMs) acquire most of their factual knowledge during the pre-training stage, through next token prediction.
arXiv:2608. 05164v1 Announce Type: cross Abstract: Independently trained large language models may develop shared internal representations of semantic concepts despite architectural differences -- but whether this geometric similarity has functional consequences for cross-model behavioural control remains untested.
arXiv:2606. 03002v1 Announce Type: cross Abstract: Quantization is a standard path to deploying large language models, and a quantized model is typically judged acceptable when its perplexity or downstream accuracy stays close to the full-precision original.
Linear probes can decode safety‑relevant concepts such as truthfulness from language‑model activations, but probe accuracy may reflect only decodability, not causal influence on model behavior. The authors show that probe weight geometry alone cannot identify the features the model actually uses, because geometrically aligned features need not be causally relevant. They introduce a sparse‑autoencoder (SAE) decomposition that ranks features by probe alignment and gradient sensitivity, and demonstrate that ablating shared, probe‑only, and random feature sets reveals a sharp dissociation: shared features drive model output changes far more than probe‑only or random features, confirming that causal relevance requires intervention beyond weight geometry.
The paper introduces the Redistribution Sensitivity Score (RSS) to detect consistent changes in sparse feature activations of Sparse Autoencoders (SAEs) when prompts are transformed. RSS identifies large, directionally consistent feature responses and evaluates their impact on downstream readouts across 27 pretrained SAEs from eight language models. Experiments show that combining RSS support erosion with the clean margin predicts recurrent safety readout failures with high AUROC, demonstrating RSS’s utility in quantifying representation changes without conflating them with failure.
arXiv:2606. 14990v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) are standard tools for mechanistic interpretability, but current SAE families are constrained by fixed encoder nonlinearities such as ReLU, JumpReLU, and TopK.