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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 1697
    Prev
    Next
    INDEX
    Explanations

    descriptions of operational activities and equipment

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    288
    +0.15
    0.7%
    566
    +0.15
    0.7%
    497
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.15
    0.07
    540
    +0.15
    0.06
    566
    +0.12
    0.03
    Negative Logits
    hus
    -0.93
    ioxide
    -0.75
    hon
    -0.72
    arton
    -0.72
    grave
    -0.71
    ilde
    -0.71
    sonian
    -0.70
     Fey
    -0.69
    username
    -0.68
    ãĤ©
    -0.68
    POSITIVE LOGITS
    eering
    0.89
    ATIONS
    0.83
     licences
    0.81
     manuals
    0.80
    ATIONAL
    0.79
    ATION
    0.79
     tempo
    0.77
    ally
    0.77
    ational
    0.77
     Procedure
    0.73
    Activations Density 0.659%

    No Known Activations