© 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. 2374
    Prev
    Next
    INDEX
    Explanations

    verbs related to undoing or preventing actions

    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
    269
    +0.13
    0.6%
    160
    +0.11
    0.5%
    195
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    269
    +0.13
    0.02
    418
    +0.11
    0.02
    9
    +0.11
    0.01
    Negative Logits
    lished
    -0.98
    Interstitial
    -0.97
    chens
    -0.89
    rics
    -0.88
    lov
    -0.77
    riers
    -0.76
    rine
    -0.76
    azines
    -0.75
    rum
    -0.74
    akov
    -0.73
    POSITIVE LOGITS
     chronological
    0.99
     engineer
    0.86
     engineered
    0.76
    engineering
    0.70
     symmetry
    0.70
     Gravity
    0.69
    kered
    0.69
     causation
    0.69
    graded
    0.69
    balanced
    0.69
    Activations Density 0.029%

    No Known Activations