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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 960
    Prev
    Next
    INDEX
    Explanations

    names of professional cyclists

    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
    520
    +0.10
    0.5%
    357
    +0.10
    0.5%
    79
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.10
    0.05
    232
    +0.10
    0.05
    579
    +0.10
    0.04
    Negative Logits
    ãģĨ
    -0.61
    ogle
    -0.61
    araoh
    -0.60
    bender
    -0.60
    SpaceEngineers
    -0.59
     Pixar
    -0.59
    ghan
    -0.58
    inki
    -0.58
     lest
    -0.57
    esame
    -0.56
    POSITIVE LOGITS
    bilt
    1.02
    emort
    0.76
    export
    0.74
    tsky
    0.67
     supra
    0.66
    rontal
    0.65
    sov
    0.64
    vich
    0.64
    ovsky
    0.63
    INAL
    0.62
    Activations Density 0.298%

    No Known Activations