© 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. 1366
    Prev
    Next
    INDEX
    Explanations

    words indicating uncertainty or doubt

    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
    331
    +0.13
    0.6%
    348
    +0.10
    0.4%
    524
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    331
    +0.13
    0.04
    348
    +0.10
    0.04
    718
    +0.09
    0.03
    Negative Logits
    tein
    -0.83
    vic
    -0.73
    ifully
    -0.71
    iful
    -0.63
    voy
    -0.63
    moil
    -0.62
    aciously
    -0.62
    vir
    -0.62
    OTO
    -0.61
    thal
    -0.61
    POSITIVE LOGITS
     anymore
    1.07
     bothered
    1.01
     bother
    0.91
     cared
    0.86
     bothering
    0.84
     mattered
    0.78
     anywhere
    0.76
    hin
    0.75
     dime
    0.74
     anything
    0.74
    Activations Density 0.075%

    No Known Activations