© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 1602
    Prev
    Next
    INDEX
    Explanations

    emotional and personal reflections on relationships, effort, and communication

    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
    288
    +0.11
    0.5%
    324
    +0.11
    0.5%
    223
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.11
    0.10
    288
    +0.11
    0.10
    324
    +0.10
    0.07
    Negative Logits
    urches
    -0.84
    itudes
    -0.80
     establishments
    -0.79
    sites
    -0.77
     Occupations
    -0.75
    tics
    -0.75
     corridors
    -0.74
     channels
    -0.73
     wards
    -0.72
     newsletters
    -0.72
    POSITIVE LOGITS
    atical
    0.90
    less
    0.85
     apiece
    0.85
    ful
    0.83
     somewhere
    0.82
     resembling
    0.81
    able
    0.81
    ogram
    0.77
     worthy
    0.74
     standpoint
    0.73
    Activations Density 0.722%

    No Known Activations