© 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. 14687
    Prev
    Next
    INDEX
    Explanations

    scientific terms related to research findings and their implications

    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
    647
    +0.10
    0.4%
    435
    +0.09
    0.4%
    134
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.10
    0.06
    435
    +0.09
    0.03
    642
    +0.09
    0.03
    Negative Logits
     whining
    -0.70
     Corinth
    -0.68
    shit
    -0.65
     kneeling
    -0.64
    lash
    -0.64
     loser
    -0.63
     POV
    -0.63
     Franch
    -0.62
     Funny
    -0.62
     Rampage
    -0.62
    POSITIVE LOGITS
     biomedical
    1.27
     scientists
    1.26
     biologists
    1.18
     neuroscience
    1.16
     scientifically
    1.14
     biomark
    1.13
     breakthrough
    1.12
     Scientists
    1.11
     someday
    1.11
     insights
    1.10
    Activations Density 0.460%

    No Known Activations