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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 16255
    Prev
    Next
    INDEX
    Explanations

    mentions of specific organizations or events by their acronym

    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
    99
    +0.17
    0.8%
    496
    +0.16
    0.7%
    659
    +0.14
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    99
    +0.17
    0.02
    659
    +0.16
    0.02
    302
    +0.14
    0.02
    Negative Logits
    loo
    -0.88
    chest
    -0.83
    tons
    -0.78
    nings
    -0.73
    thing
    -0.72
    gerald
    -0.71
    shake
    -0.66
    butt
    -0.66
    tery
    -0.65
     Aires
    -0.63
    POSITIVE LOGITS
    ODE
    1.03
    urses
    1.03
    STAR
    0.95
    ategor
    0.91
    ERT
    0.82
    ossus
    0.82
    ORE
    0.82
    OUNT
    0.81
    ouri
    0.80
    UFC
    0.78
    Activations Density 0.014%

    No Known Activations