© 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. 17222
    Prev
    Next
    INDEX
    Explanations

    negative sentiment or contradictions

    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
    373
    +0.13
    0.6%
    377
    +0.12
    0.6%
    437
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    437
    +0.13
    0.04
    226
    +0.12
    0.04
    449
    +0.10
    0.04
    Negative Logits
    kamp
    -0.73
    rift
    -0.72
    ãĤ¼
    -0.65
    istle
    -0.65
    ç·
    -0.64
     Mehran
    -0.63
    USH
    -0.62
    ãĥį
    -0.59
    usher
    -0.59
    yout
    -0.59
    POSITIVE LOGITS
    epad
    1.10
    hin
    1.03
     necessarily
    1.01
    icably
    0.95
    withstanding
    0.94
     bothering
    0.94
    eworthy
    0.93
    icable
    0.92
    icing
    0.89
    ices
    0.88
    Activations Density 0.063%

    No Known Activations