Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 19760
    Prev
    Next
    INDEX
    Explanations

    words related to holding on to beliefs or physical objects

    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
    587
    +0.11
    0.5%
    3
    +0.10
    0.5%
    535
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    587
    +0.11
    0.02
    3
    +0.10
    0.02
    761
    +0.10
    0.02
    Negative Logits
    eger
    -0.77
    iterranean
    -0.77
    £ı
    -0.75
    ftime
    -0.71
    Ĥª
    -0.69
    imaru
    -0.68
    arthed
    -0.68
    omsky
    -0.67
    cius
    -0.66
     SPORTS
    -0.64
    POSITIVE LOGITS
     tight
    0.89
     tighter
    0.85
     reins
    0.78
    hold
    0.75
     virginity
    0.73
     dear
    0.72
     hope
    0.71
     Tight
    0.70
     ransom
    0.69
    oil
    0.68
    Activations Density 0.046%

    No Known Activations