© 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. 10763
    Prev
    Next
    INDEX
    Explanations

    references to domain management and DNS settings

    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
    +0.14
    0.6%
    137
    +0.12
    0.5%
    430
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.03
    137
    +0.12
    0.01
    403
    +0.11
    0.01
    Negative Logits
    tics
    -0.79
     Clement
    -0.78
    eph
    -0.76
    thus
    -0.75
    pter
    -0.75
    hower
    -0.73
    iah
    -0.71
    aquin
    -0.69
    zz
    -0.68
    eh
    -0.67
    POSITIVE LOGITS
     registrations
    0.82
     domains
    0.81
     regist
    0.80
     suffix
    0.76
     domain
    0.75
     name
    0.75
     names
    0.74
    Market
    0.73
     ont
    0.72
     lookup
    0.71
    Activations Density 0.088%

    No Known Activations