© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 1434
    Prev
    Next
    INDEX
    Explanations

    words related to specific locations, particularly "Sierra."

    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
    659
    +0.11
    0.5%
    728
    +0.11
    0.5%
    18
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.11
    0.03
    288
    +0.11
    0.03
    232
    +0.11
    0.03
    Negative Logits
    ifier
    -0.85
    sworth
    -0.81
    esty
    -0.73
    ipal
    -0.72
    race
    -0.71
    assic
    -0.70
    ors
    -0.70
    estate
    -0.70
    ewski
    -0.68
    idy
    -0.68
    POSITIVE LOGITS
     Leone
    1.42
     Nevada
    1.00
     Scotia
    0.96
     Alto
    0.84
     Isles
    0.83
     Guinea
    0.79
     Vista
    0.77
     Lauder
    0.77
    vernment
    0.76
    llan
    0.76
    Activations Density 0.126%

    No Known Activations