© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 12512
    Prev
    Next
    INDEX
    Explanations

    the name "Maddow," potentially related to Rachel Maddow, repeatedly throughout the text segment

    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
    526
    +0.16
    0.7%
    538
    +0.13
    0.6%
    343
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    343
    +0.16
    0.01
    161
    +0.13
    0.01
    538
    +0.13
    0.01
    Negative Logits
    âĸ¬
    -0.80
     Origin
    -0.71
     microwave
    -0.71
    PDATE
    -0.70
    ccording
    -0.66
    0000000000000000
    -0.65
     sampled
    -0.65
     lunch
    -0.63
    ASED
    -0.63
     ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³
    -0.61
    POSITIVE LOGITS
    ings
    0.98
     Madd
    0.93
    ipedia
    0.92
    itude
    0.92
    erton
    0.91
    owitz
    0.89
    ox
    0.89
    ocks
    0.88
    ow
    0.88
    ington
    0.86
    Activations Density 0.019%

    No Known Activations