© 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 EvalsExportsinterp-engineNEWAPI 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. 2442
    Prev
    Next
    INDEX
    Explanations

    instances of the word "that" followed by a strong activation

    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
    393
    +0.13
    0.6%
    382
    +0.11
    0.5%
    274
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    601
    +0.13
    0.05
    562
    +0.11
    0.05
    203
    +0.10
    0.05
    Negative Logits
    hens
    -0.81
    oran
    -0.71
    obb
    -0.70
    ãĥ¼ãĤ¯
    -0.69
    tails
    -0.68
    arest
    -0.68
    uty
    -0.67
    raq
    -0.67
    ostic
    -0.67
    oufl
    -0.67
    POSITIVE LOGITS
     they
    0.91
     someday
    0.83
     pesky
    0.83
     there
    0.81
     THEY
    0.81
     fateful
    0.79
     although
    0.75
     we
    0.73
     THERE
    0.70
    genre
    0.69
    Activations Density 0.236%

    No Known Activations