© 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 12288
    9. 8-RES_FS12288-JB
    11. 10705
    Prev
    Next
    INDEX
    Explanations

    references to sections or parts of a document

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    306
    +0.11
    0.5%
    604
    +0.11
    0.5%
    395
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    306
    +0.11
    0.03
    604
    +0.11
    0.02
    395
    +0.10
    0.02
    Negative Logits
    cffffcc
    -0.80
    bably
    -0.79
    biased
    -0.77
    operated
    -0.76
    icum
    -0.76
    affle
    -0.75
    âĹ¼
    -0.73
    aced
    -0.73
    purpose
    -0.72
    iated
    -0.71
    POSITIVE LOGITS
     below
    1.03
     supra
    0.95
     Also
    0.86
     sidebar
    0.81
     above
    0.78
     Below
    0.76
     TIME
    0.75
     ALSO
    0.74
     Above
    0.73
    KER
    0.72
    Activations Density 0.047%

    No Known Activations