© 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. 8-RES_POST_32K-OAI
    11. 6342
    Prev
    Next
    INDEX
    Explanations

    common abbreviations and shorthand terms used in casual text

    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
    184
    +0.11
    0.5%
    354
    +0.09
    0.4%
    585
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.11
    0.02
    701
    +0.09
    0.02
    406
    +0.09
    0.02
    Negative Logits
    ãĥ¼ãĥĨ
    -0.66
    issors
    -0.63
     Canaver
    -0.59
    gypt
    -0.59
    ̶
    -0.59
    aign
    -0.58
     ru
    -0.57
    furt
    -0.56
     cord
    -0.56
     withstand
    -0.55
    POSITIVE LOGITS
    urnal
    1.29
    mber
    0.85
    ODY
    0.79
     Chomsky
    0.76
    EMBER
    0.73
    YP
    0.73
    omi
    0.71
    ilus
    0.67
    agon
    0.67
    ourse
    0.66
    Activations Density 0.118%

    No Known Activations