© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 8799
    Prev
    Next
    INDEX
    Explanations

    instances of the word "publicly" and related terms discussing public visibility or disclosure

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    503
    +0.12
    0.6%
    554
    +0.11
    0.5%
    532
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    532
    +0.12
    0.02
    554
    +0.11
    0.01
    151
    +0.10
    0.03
    Negative Logits
     Procedure
    -0.79
    ragon
    -0.78
     Clever
    -0.75
    wolves
    -0.74
    nian
    -0.74
     Grimoire
    -0.72
     Rover
    -0.72
    mble
    -0.71
    nesota
    -0.69
     Cameroon
    -0.68
    POSITIVE LOGITS
     humiliated
    0.95
     shaming
    0.90
     acknowledged
    0.89
     proclaimed
    0.82
     denounced
    0.82
     humiliating
    0.81
     humili
    0.80
     denouncing
    0.80
     apologize
    0.78
     endorsed
    0.77
    Activations Density 0.054%

    No Known Activations