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. 1498
    Prev
    Next
    INDEX
    Explanations

    news or information related to current events and politics

    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
    87
    +0.45
    4.4%
    64
    +0.07
    0.7%
    756
    +0.06
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.45
    -0.09
    756
    +0.07
    -0.16
    570
    +0.06
    -0.06
    Negative Logits
    ".[
    -0.32
     attRot
    -0.31
    .ãĢį
    -0.29
     [...]
    -0.28
     ....
    -0.28
    )...
    -0.28
    !".
    -0.27
     Allaah
    -0.27
     [â̦]
    -0.27
    Âł
    -0.27
    POSITIVE LOGITS
    umably
    0.31
    TPPStreamerBot
    0.29
    earcher
    0.28
    lier
    0.27
    erential
    0.27
    itely
    0.27
    mentioned
    0.27
    lightly
    0.26
    Ĭ±
    0.26
    ilant
    0.26
    Activations Density 17.394%

    No Known Activations