© 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. 19363
    Prev
    Next
    INDEX
    Explanations

    file paths and commands in a computer system

    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
    373
    +0.12
    0.5%
    160
    +0.11
    0.5%
    700
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.12
    0.02
    107
    +0.11
    0.02
    700
    +0.11
    0.02
    Negative Logits
     Beir
    -0.84
    terday
    -0.71
     Sie
    -0.69
     Miche
    -0.68
     Faw
    -0.68
     Faul
    -0.67
     Schultz
    -0.67
     Labrador
    -0.67
     undert
    -0.66
     Quest
    -0.65
    POSITIVE LOGITS
    usr
    1.41
    tmp
    1.37
    Users
    1.31
    home
    1.26
    Applications
    1.17
    api
    1.14
    opt
    1.12
    wcsstore
    1.10
    proc
    1.09
    :/
    1.06
    Activations Density 0.028%

    No Known Activations