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 1536
    5. 8-RES_FS1536-JB
    6. 222
    Prev
    Next
    INDEX
    Explanations

    technical and medical terms, particularly those related to medical services and emergencies

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs1536-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    1,536
    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
    755
    +0.10
    0.5%
    13
    +0.10
    0.4%
    422
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.13
    326
    +0.10
    0.10
    346
    +0.09
    0.09
    Negative Logits
     Cumber
    -0.85
    eral
    -0.71
    ãĥ¤
    -0.71
    essee
    -0.68
    erences
    -0.67
    ricting
    -0.66
    urion
    -0.66
    hesis
    -0.61
    _>
    -0.60
    aido
    -0.60
    POSITIVE LOGITS
    wings
    1.00
    planes
    0.91
    plane
    0.89
    walk
    0.88
     Flight
    0.87
     hangar
    0.85
     flights
    0.83
     Airlines
    0.83
    hawk
    0.82
     flying
    0.82
    Activations Density 2.309%

    No Known Activations