© 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. 7715
    Prev
    Next
    INDEX
    Explanations

    announcements and updates about events or projects

    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
    120
    +0.12
    0.5%
    373
    +0.11
    0.5%
    265
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    267
    +0.12
    0.07
    598
    +0.11
    0.05
    149
    +0.11
    0.04
    Negative Logits
     disputed
    -0.75
     discredited
    -0.69
     repud
    -0.67
     eroded
    -0.67
     discrimination
    -0.66
    SPONSORED
    -0.66
    Prosecutors
    -0.63
     dispro
    -0.63
     rejected
    -0.62
     coercive
    -0.62
    POSITIVE LOGITS
     THANK
    1.07
    Enjoy
    0.91
     PLEASE
    0.90
     thank
    0.89
     Patreon
    0.88
     Tune
    0.88
    Alright
    0.87
     congr
    0.87
     ðŁĻĤ
    0.85
     please
    0.84
    Activations Density 0.611%

    No Known Activations