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. 21477
    Prev
    Next
    INDEX
    Explanations

    data related to government policies and statistics

    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
    223
    +0.12
    0.5%
    199
    +0.12
    0.5%
    731
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    223
    +0.12
    0.06
    199
    +0.12
    0.05
    731
    +0.11
    0.04
    Negative Logits
     histories
    -0.63
     mascul
    -0.61
    enges
    -0.60
    birds
    -0.57
    ypes
    -0.57
     Lumpur
    -0.57
     angles
    -0.57
    orative
    -0.56
     Pompe
    -0.56
    folk
    -0.56
    POSITIVE LOGITS
    st
    1.33
    Password
    1.32
    125
    1.27
    120
    1.25
    ½
    1.07
    128
    1.05
    000000
    1.04
    123
    0.97
    âģ
    0.97
    0000000
    0.87
    Activations Density 0.091%

    No Known Activations