© 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. 19399
    Prev
    Next
    INDEX
    Explanations

    dates and descriptions of events leading up to or following certain actions or situations

    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
    430
    +0.12
    0.5%
    286
    +0.10
    0.5%
    147
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    286
    +0.12
    0.04
    59
    +0.10
    0.03
    476
    +0.10
    0.03
    Negative Logits
    geist
    -0.81
    hack
    -0.74
    ische
    -0.71
    ISM
    -0.70
    Force
    -0.69
    Fight
    -0.69
    ophe
    -0.68
    arna
    -0.67
    aez
    -0.66
    WIND
    -0.65
    POSITIVE LOGITS
     settling
    1.15
     returning
    1.08
     retiring
    1.05
     realizing
    1.04
     disappearing
    1.03
     committing
    1.00
     concluding
    1.00
     deciding
    0.97
     abruptly
    0.96
     reaching
    0.96
    Activations Density 0.059%

    No Known Activations