WriteSAE is the first sparse autoencoder that factors decoder atoms into the native d_k x d_v cache write shape of recurrent models and supplies a closed-form per-token logit shift for atom substitution.
hub
Finding neurons in a haystack: Case studies with sparse probing
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
hub tools
representative citing papers
AVLLMs encode audio semantics in middle layers but suppress them in final text outputs when audio conflicts with vision, due to training that largely inherits from vision-language base models.
K-sparse autoencoders with dead-latent fixes produce clean scaling laws and better feature quality metrics that improve with size, shown by training a 16-million-latent model on GPT-4 activations.
Suppressing one refusal neuron or amplifying one concept neuron bypasses safety alignment in LLMs from 1.7B to 70B parameters without training or prompt engineering.
Tree SAE learns hierarchical feature structures by combining activation coverage with a new reconstruction condition, outperforming prior SAEs on hierarchical pair detection while matching state-of-the-art benchmark performance.
Perturbation probing identifies tiny sets of FFN neurons that control refusal templates and language routing in LLMs, enabling precise ablations and directional interventions that alter behavior on benchmarks while preserving safety.
LIRA aligns latent instruction representations in LLMs to defend against jailbreaks, backdoors, and undesired knowledge, blocking over 99% of PEZ attacks and achieving optimal WMDP forgetting.
Pretrained vision transformers use specific attention heads sensitive to Gestalt continuity for object binding, shown via probes on synthetic datasets and ablation experiments.
GPT-2 Small's terminal MLP implements a legible three-tier exception handler with 27 named neurons that routes predictions, while previously identified knowledge neurons function as amplifiers of residual-stream signals rather than fact storage.
A mechanics of the learning process is emerging in deep learning theory, characterized by dynamics, coarse statistics, and falsifiable predictions across idealized settings, limits, laws, hyperparameters, and universal behaviors.
citing papers explorer
-
WriteSAE: Sparse Autoencoders for Recurrent State
WriteSAE is the first sparse autoencoder that factors decoder atoms into the native d_k x d_v cache write shape of recurrent models and supplies a closed-form per-token logit shift for atom substitution.
-
Do Audio-Visual Large Language Models Really See and Hear?
AVLLMs encode audio semantics in middle layers but suppress them in final text outputs when audio conflicts with vision, due to training that largely inherits from vision-language base models.
-
Scaling and evaluating sparse autoencoders
K-sparse autoencoders with dead-latent fixes produce clean scaling laws and better feature quality metrics that improve with size, shown by training a 16-million-latent model on GPT-4 activations.
-
A Single Neuron Is Sufficient to Bypass Safety Alignment in Large Language Models
Suppressing one refusal neuron or amplifying one concept neuron bypasses safety alignment in LLMs from 1.7B to 70B parameters without training or prompt engineering.
-
Tree SAE: Learning Hierarchical Feature Structures in Sparse Autoencoders
Tree SAE learns hierarchical feature structures by combining activation coverage with a new reconstruction condition, outperforming prior SAEs on hierarchical pair detection while matching state-of-the-art benchmark performance.
-
Perturbation Probing: A Two-Pass-per-Prompt Diagnostic for FFN Behavioral Circuits in Aligned LLMs
Perturbation probing identifies tiny sets of FFN neurons that control refusal templates and language routing in LLMs, enabling precise ablations and directional interventions that alter behavior on benchmarks while preserving safety.
-
Latent Instruction Representation Alignment: defending against jailbreaks, backdoors and undesired knowledge in LLMs
LIRA aligns latent instruction representations in LLMs to defend against jailbreaks, backdoors, and undesired knowledge, blocking over 99% of PEZ attacks and achieving optimal WMDP forgetting.
-
I Walk the Line: Examining the Role of Gestalt Continuity in Object Binding for Vision Transformers
Pretrained vision transformers use specific attention heads sensitive to Gestalt continuity for object binding, shown via probes on synthetic datasets and ablation experiments.
-
Darkness Visible: Reading the Exception Handler of a Language Model
GPT-2 Small's terminal MLP implements a legible three-tier exception handler with 27 named neurons that routes predictions, while previously identified knowledge neurons function as amplifiers of residual-stream signals rather than fact storage.
-
There Will Be a Scientific Theory of Deep Learning
A mechanics of the learning process is emerging in deep learning theory, characterized by dynamics, coarse statistics, and falsifiable predictions across idealized settings, limits, laws, hyperparameters, and universal behaviors.