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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 981
    Prev
    Next
    INDEX
    Explanations

    mentions of Facebook

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    247
    +0.13
    0.6%
    343
    +0.12
    0.5%
    585
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    247
    +0.13
    0.04
    609
    +0.12
    0.04
    343
    +0.11
    0.04
    Negative Logits
    nces
    -0.82
    DERR
    -0.69
    heny
    -0.69
    cised
    -0.68
    ved
    -0.68
    rosc
    -0.68
    inence
    -0.67
    zin
    -0.67
    lass
    -0.66
    ptin
    -0.66
    POSITIVE LOGITS
     Messenger
    0.99
    nect
    0.95
    username
    0.88
     Live
    0.85
     Pages
    0.82
    acebook
    0.78
    osphere
    0.77
     Timeline
    0.76
     Caf
    0.75
     Username
    0.74
    Activations Density 0.064%

    No Known Activations