© 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 12288
    9. 8-RES_FS12288-JB
    11. 9389
    Prev
    Next
    INDEX
    Explanations

    names of TV show hosts

    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
    757
    +0.13
    0.6%
    718
    +0.10
    0.5%
    578
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.13
    0.05
    232
    +0.10
    0.04
    659
    +0.10
    0.04
    Negative Logits
     redes
    -0.67
    âĸ¬
    -0.66
     Oval
    -0.66
    elist
    -0.65
    efeated
    -0.63
     Izan
    -0.63
    onym
    -0.60
    -+-+
    -0.60
    nered
    -0.60
    ĺħ
    -0.59
    POSITIVE LOGITS
    nuts
    0.98
    nut
    0.91
    adow
    0.91
    ty
    0.89
    merga
    0.84
    ishes
    0.80
    ishi
    0.80
    agin
    0.79
    awan
    0.75
     Kimmel
    0.75
    Activations Density 0.214%

    No Known Activations