© 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. 10720
    Prev
    Next
    INDEX
    Explanations

    links or URLs to websites and online content

    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
    402
    +0.13
    0.6%
    565
    +0.12
    0.6%
    488
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.13
    0.04
    488
    +0.12
    0.04
    462
    +0.11
    0.03
    Negative Logits
    TPS
    -0.86
    Spread
    -0.77
    âĸ¬
    -0.75
    WAYS
    -0.74
    Enough
    -0.71
    Tracker
    -0.71
    TIME
    -0.70
    SHIP
    -0.69
    MIC
    -0.68
    Ins
    -0.68
    POSITIVE LOGITS
    ffe
    1.05
    fe
    1.04
    cf
    0.94
    ae
    0.93
    efe
    0.93
    fc
    0.89
    cd
    0.87
    afa
    0.86
    fd
    0.85
    fff
    0.83
    Activations Density 0.113%

    No Known Activations