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 768
    5. 8-RES_FS768-JB
    6. 486
    Prev
    Next
    INDEX
    Explanations

    countries, nationalities, and related geopolitical terms

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    267
    +0.12
    0.5%
    288
    +0.12
    0.5%
    57
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    752
    +0.12
    0.17
    57
    +0.12
    0.13
    288
    +0.11
    0.15
    Negative Logits
    utherford
    -0.87
    anship
    -0.81
    eus
    -0.80
    ueller
    -0.79
    nels
    -0.78
    ording
    -0.78
    ertodd
    -0.78
    chery
    -0.77
    mble
    -0.77
    elling
    -0.77
    POSITIVE LOGITS
     Nadu
    1.08
     nationals
    1.08
    apolis
    1.04
     Federation
    0.96
    istani
    0.95
     Union
    0.95
     citizen
    0.94
     Embassy
    0.94
     citizens
    0.94
     Airlines
    0.93
    Activations Density 2.025%

    No Known Activations