© 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 24576
    9. 8-RES_FS24576-JB
    11. 9684
    Prev
    Next
    INDEX
    Explanations

    keywords related to technical instructions or guidance, such as programming languages, components, and step-by-step procedures

    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
    481
    +0.20
    0.9%
    447
    +0.17
    0.8%
    679
    +0.14
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    679
    +0.20
    0.07
    455
    +0.17
    0.06
    459
    +0.14
    0.05
    Negative Logits
    atever
    -0.64
    laughs
    -0.58
     Curious
    -0.58
    ãģĻ
    -0.57
    idan
    -0.57
    externalActionCode
    -0.56
    hemy
    -0.56
     Pillar
    -0.56
    ference
    -0.56
     eternity
    -0.55
    POSITIVE LOGITS
     are
    1.02
     have
    0.95
     require
    0.95
     involve
    0.92
     were
    0.91
     contain
    0.89
     sprang
    0.86
     fared
    0.85
     seem
    0.84
     originate
    0.84
    Activations Density 0.274%

    No Known Activations