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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 45746
    Prev
    Next
    INDEX
    Explanations

    references to authority and organizational structures

    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
    314
    +0.13
    0.6%
    513
    +0.12
    0.5%
    525
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.04
    205
    +0.12
    0.03
    219
    +0.11
    0.03
    Negative Logits
    ampires
    -0.68
    wered
    -0.67
    aughs
    -0.67
     whiff
    -0.65
     slightest
    -0.65
    adish
    -0.64
    Downloadha
    -0.64
     realistically
    -0.60
     lowly
    -0.60
     candles
    -0.58
    POSITIVE LOGITS
    igsaw
    0.86
     repertoire
    0.80
     continuum
    0.78
     endeavour
    0.76
     regimen
    0.73
     tradition
    0.70
     package
    0.70
     constellation
    0.70
     puzzle
    0.69
    ĸļ
    0.69
    Activations Density 0.190%

    No Known Activations