© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 10-RES_POST_32K-OAI
    11. 30902
    Prev
    Next
    INDEX
    Explanations

    punctuation and sentence boundaries

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    6
    +0.12
    0.6%
    669
    +0.11
    0.5%
    361
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.12
    0.26
    107
    +0.11
    0.24
    102
    +0.10
    0.20
    Negative Logits
     advoc
    -0.92
    yip
    -0.87
     stated
    -0.83
     stating
    -0.78
    catentry
    -0.78
     confir
    -0.78
     util
    -0.76
    wiki
    -0.75
    rontal
    -0.75
     concess
    -0.75
    POSITIVE LOGITS
    toggle
    1.29
    Enlarge
    1.22
    advertisement
    1.21
    Newsletter
    1.21
    Yet
    1.10
    Then
    1.02
    Advertisement
    1.01
    Suddenly
    0.98
    âĢ¢âĢ¢
    0.95
     Everywhere
    0.95
    Activations Density 0.856%

    No Known Activations