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 3072
    5. 8-RES_FS3072-JB
    6. 1904
    Prev
    Next
    INDEX
    Explanations

    scientific terms and concepts related to geology and chemistry

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    586
    +0.10
    0.5%
    651
    +0.10
    0.5%
    235
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    539
    +0.10
    0.07
    235
    +0.10
    0.08
    290
    +0.09
    0.06
    Negative Logits
    iquette
    -0.85
    estate
    -0.85
    deal
    -0.74
    atism
    -0.73
     Corinth
    -0.72
     kneeling
    -0.72
     grievances
    -0.72
     whine
    -0.71
    TRUMP
    -0.70
     hating
    -0.69
    POSITIVE LOGITS
     researchers
    1.61
     scientists
    1.50
    Researchers
    1.47
     Researchers
    1.42
     biologists
    1.35
     researcher
    1.32
    Scientists
    1.31
     Scientists
    1.31
     experiments
    1.30
     laboratory
    1.30
    Activations Density 2.477%

    No Known Activations