© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 4042
    Prev
    Next
    INDEX
    Explanations

    words related to safety and concerns regarding potential dangers or risks

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    679
    +0.11
    0.5%
    52
    +0.10
    0.4%
    336
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    86
    +0.11
    0.03
    52
    +0.10
    0.03
    51
    +0.10
    0.03
    Negative Logits
    ingers
    -0.76
    lite
    -0.71
     Carbuncle
    -0.63
    nice
    -0.63
     Bom
    -0.61
     Swords
    -0.59
    Completed
    -0.59
    avour
    -0.59
     swe
    -0.58
    ann
    -0.57
    POSITIVE LOGITS
    warts
    1.12
    wart
    1.06
    lessly
    0.93
     trolling
    0.87
    ingly
    0.82
     raised
    0.79
     regarding
    0.75
     voiced
    0.73
     concerns
    0.72
    edly
    0.71
    Activations Density 0.038%

    No Known Activations