© 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 49152
    9. 8-RES_FS49152-JB
    11. 26536
    Prev
    Next
    INDEX
    Explanations

    concepts related to intelligence and oversight within bureaucracies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    160
    +0.10
    0.4%
    248
    +0.09
    0.4%
    548
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    4
    +0.10
    0.01
    404
    +0.09
    0.01
    248
    +0.09
    0.01
    Negative Logits
    atron
    -0.72
     Aires
    -0.69
     Jagu
    -0.67
     simulate
    -0.67
     Moroc
    -0.67
     casing
    -0.65
     Leonardo
    -0.62
    ilan
    -0.61
     forfeit
    -0.61
    alys
    -0.61
    POSITIVE LOGITS
    ÂŃ
    1.03
     Dems
    0.89
    .—
    0.86
     PACs
    0.80
    ptroller
    0.80
     Democrats
    0.78
    lihood
    0.77
     Republicans
    0.77
     Senate
    0.77
     Democr
    0.76
    Activations Density 0.023%

    No Known Activations