© 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. 22409
    Prev
    Next
    INDEX
    Explanations

    phrases and words associated with training and education

    oai_token-act-pair · gpt-4o-miniTriggered 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
    575
    +0.15
    0.7%
    664
    +0.09
    0.4%
    288
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.15
    0.03
    469
    +0.09
    0.02
    664
    +0.09
    0.02
    Negative Logits
     inconven
    -0.69
     Amen
    -0.65
     gloom
    -0.59
     inconvenience
    -0.59
     goodbye
    -0.59
    osi
    -0.58
    ogene
    -0.58
     scattering
    -0.58
     attributable
    -0.58
     Rodham
    -0.58
    POSITIVE LOGITS
    romy
    0.88
    arnaev
    0.86
    learn
    0.82
     athletes
    0.81
     apprentices
    0.79
     instructors
    0.78
    arse
    0.77
    dinand
    0.77
     Swordsman
    0.77
    anooga
    0.75
    Activations Density 0.045%

    No Known Activations