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. 780
    Prev
    Next
    INDEX
    Explanations

    Washington, D.C

    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
    254
    +0.14
    0.6%
    23
    +0.11
    0.5%
    546
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    254
    +0.14
    0.04
    200
    +0.11
    0.04
    752
    +0.10
    0.04
    Negative Logits
    vana
    -0.80
    leness
    -0.76
    ONY
    -0.73
    otal
    -0.64
    selection
    -0.63
    ItemImage
    -0.63
    flush
    -0.63
     Cameroon
    -0.63
    lihood
    -0.62
    ensed
    -0.62
    POSITIVE LOGITS
     DC
    1.17
    sonian
    1.09
     Post
    1.08
     Capitals
    1.03
     Heights
    1.02
     Dull
    1.01
     Redskins
    0.94
     Examiner
    0.94
     Irving
    0.94
     Nationals
    0.91
    Activations Density 0.030%

    No Known Activations