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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 10786
    Prev
    Next
    INDEX
    Explanations

    words related to potential risks or threats

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    447
    +0.12
    0.5%
    551
    +0.11
    0.5%
    438
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    438
    +0.12
    0.03
    271
    +0.11
    0.03
    34
    +0.10
    0.03
    Negative Logits
    guyen
    -0.79
    owned
    -0.77
    elle
    -0.77
    olitan
    -0.74
    hew
    -0.73
    Ħ¢
    -0.72
    pel
    -0.71
    issance
    -0.71
    ļéĨĴ
    -0.70
    yard
    -0.69
    POSITIVE LOGITS
     Danger
    0.86
     endanger
    0.82
     dangers
    0.80
     lurking
    0.79
    danger
    0.78
     danger
    0.77
     lur
    0.76
    ously
    0.75
     slope
    0.75
     threats
    0.74
    Activations Density 0.097%

    No Known Activations