© 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. 4778
    Prev
    Next
    INDEX
    Explanations

    phrases related to controversial or impactful events

    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
    402
    +0.12
    0.5%
    757
    +0.11
    0.5%
    65
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    402
    +0.12
    0.04
    355
    +0.11
    0.04
    65
    +0.11
    0.03
    Negative Logits
    izons
    -0.95
    azines
    -0.72
    ulner
    -0.67
    ycle
    -0.66
     brakes
    -0.66
    omever
    -0.65
    iors
    -0.65
    anyon
    -0.64
    instead
    -0.64
    usercontent
    -0.64
    POSITIVE LOGITS
     nutshell
    1.06
    Synopsis
    0.87
     reversal
    0.85
    catentry
    0.79
     understatement
    0.79
     ironic
    0.78
     twist
    0.77
     irony
    0.75
    stakes
    0.75
     coincidence
    0.75
    Activations Density 0.159%

    No Known Activations