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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    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. 11803
    Prev
    Next
    INDEX
    Explanations

    phrases related to personal opinions or preferences

    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
    604
    +0.12
    0.5%
    747
    +0.11
    0.5%
    52
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    604
    +0.12
    0.04
    747
    +0.11
    0.04
    52
    +0.10
    0.04
    Negative Logits
    idium
    -0.70
     concess
    -0.70
     Mandatory
    -0.61
    isoft
    -0.60
    ivari
    -0.60
    puff
    -0.58
    rolet
    -0.58
    ignore
    -0.58
    cussion
    -0.57
    oiler
    -0.57
    POSITIVE LOGITS
     fault
    0.80
    ById
    0.76
    ibility
    0.74
     shelter
    0.73
     irresistible
    0.72
     inspiration
    0.72
    ered
    0.71
     refuge
    0.71
     amusement
    0.71
     satisfaction
    0.70
    Activations Density 0.063%

    No Known Activations