© 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. 17277
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal or official swearings-in

    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
    297
    +0.09
    0.4%
    85
    +0.09
    0.4%
    289
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    85
    +0.09
    0.01
    297
    +0.09
    0.01
    37
    +0.09
    0.01
    Negative Logits
     guiActiveUnfocused
    -0.79
    bnb
    -0.77
     GOODMAN
    -0.72
    çĶ°
    -0.72
     Shutterstock
    -0.68
    neau
    -0.68
    nesota
    -0.67
    Stage
    -0.65
    markets
    -0.64
    SET
    -0.64
    POSITIVE LOGITS
     oath
    1.22
     sworn
    1.10
     swearing
    0.94
     allegiance
    0.91
     swore
    0.88
    iffs
    0.87
     vow
    0.87
     vengeance
    0.82
    atory
    0.81
     duty
    0.79
    Activations Density 0.013%

    No Known Activations