© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 24256
    Prev
    Next
    INDEX
    Explanations

    keywords related to specific topics or concepts that indicate a significant focus area in the text

    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
    314
    +0.14
    0.7%
    288
    +0.12
    0.6%
    695
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.09
    575
    +0.12
    0.07
    524
    +0.10
    0.05
    Negative Logits
    lees
    -0.85
    ernels
    -0.84
    Americans
    -0.78
    usions
    -0.78
    ouls
    -0.75
    aneously
    -0.75
    awatts
    -0.73
    itudes
    -0.72
    tails
    -0.72
    liners
    -0.72
    POSITIVE LOGITS
     standpoint
    1.08
     program
    0.92
     programme
    0.88
     aide
    0.86
     advisory
    0.85
     department
    0.85
     counselor
    0.84
     facility
    0.84
     regimen
    0.83
     outlet
    0.81
    Activations Density 0.351%

    No Known Activations