© 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. 18463
    Prev
    Next
    INDEX
    Explanations

    conversations in a customer service or technical support setting

    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
    496
    +0.14
    0.7%
    228
    +0.11
    0.5%
    409
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    314
    +0.14
    0.01
    232
    +0.11
    0.04
    503
    +0.11
    0.03
    Negative Logits
     Kaf
    -0.72
     Downs
    -0.69
     Kits
    -0.65
     ham
    -0.65
     ......
    -0.64
     Austral
    -0.64
     [&
    -0.64
     Hut
    -0.63
     Pigs
    -0.62
     di
    -0.62
    POSITIVE LOGITS
    ellectual
    0.90
    yright
    0.90
    arted
    0.86
    romeda
    0.84
    emption
    0.83
    resa
    0.81
    hing
    0.81
    etition
    0.81
    rust
    0.80
    amina
    0.79
    Activations Density 0.117%

    No Known Activations