© 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. 22096
    Prev
    Next
    INDEX
    Explanations

    programming commands and flags related to installation and compilation

    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
    637
    +0.13
    0.6%
    453
    +0.09
    0.4%
    700
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    637
    +0.13
    0.02
    326
    +0.09
    0.01
    700
    +0.09
    0.02
    Negative Logits
    hower
    -0.66
    ipedia
    -0.64
    kaya
    -0.64
    anchester
    -0.63
    iaries
    -0.62
     jams
    -0.61
    eria
    -0.60
    hess
    -0.60
    ifice
    -0.59
     Reborn
    -0.59
    POSITIVE LOGITS
    /-
    0.89
    /+
    0.69
     ~/.
    0.68
    ailable
    0.66
    Static
    0.66
    qq
    0.62
     (>
    0.62
     *.
    0.61
     "$
    0.61
     privileged
    0.60
    Activations Density 0.036%

    No Known Activations