© 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 24576
    9. 8-RES_FS24576-JB
    11. 9838
    Prev
    Next
    INDEX
    Explanations

    proper nouns and names of individuals 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_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
    659
    +0.12
    0.5%
    718
    +0.11
    0.5%
    159
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.12
    0.05
    478
    +0.11
    0.05
    263
    +0.09
    0.05
    Negative Logits
     Noir
    -0.74
     bottleneck
    -0.67
     hust
    -0.64
     CPC
    -0.61
     bars
    -0.60
     jammed
    -0.60
     remarks
    -0.59
     CDs
    -0.58
     jams
    -0.58
     ashore
    -0.58
    POSITIVE LOGITS
    prising
    1.33
    nexpected
    1.25
    seless
    1.19
    lyss
    1.19
    PDATED
    1.17
    gly
    1.11
    pperc
    1.10
    topia
    1.08
    mpire
    1.03
    NA
    0.98
    Activations Density 0.055%

    No Known Activations