© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 22134
    Prev
    Next
    INDEX
    Explanations

    information related to tiles used in construction or flooring

    oai_token-act-pair · gpt-4o-miniTriggered 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
    288
    +0.14
    0.6%
    0
    +0.12
    0.6%
    289
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.02
    326
    +0.12
    0.01
    289
    +0.10
    0.01
    Negative Logits
    ulner
    -0.95
    ournal
    -0.79
    >[
    -0.69
    arcer
    -0.68
    udence
    -0.68
    ickson
    -0.67
    hovah
    -0.66
    ĨĴ
    -0.66
     largeDownload
    -0.65
     vulner
    -0.64
    POSITIVE LOGITS
     tiles
    1.10
     tile
    0.98
     Tile
    0.94
    sheet
    0.89
     grids
    0.85
     sheets
    0.85
     roofs
    0.84
    stone
    0.83
    ade
    0.81
    tery
    0.80
    Activations Density 0.028%

    No Known Activations