© 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. 10353
    Prev
    Next
    INDEX
    Explanations

    information related to grants, programs, and funding in institutions or organizations

    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
    362
    +0.16
    0.8%
    92
    +0.11
    0.5%
    710
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.16
    0.09
    92
    +0.11
    0.06
    608
    +0.11
    0.07
    Negative Logits
    mire
    -0.88
    ĸļ
    -0.87
    mill
    -0.80
    ocument
    -0.80
    onda
    -0.79
    irez
    -0.77
    elected
    -0.75
    hess
    -0.74
    ":[
    -0.73
    gate
    -0.73
    POSITIVE LOGITS
     own
    1.55
     chances
    1.27
     ability
    1.20
     reputation
    1.15
     relationship
    1.10
     fortunes
    1.08
     competitiveness
    1.06
     commitment
    1.05
     repertoire
    1.05
     finances
    1.04
    Activations Density 0.322%

    No Known Activations