© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 15518
    Prev
    Next
    INDEX
    Explanations

    phrases signaling a transition or change from one state to another

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    608
    +0.16
    0.7%
    425
    +0.10
    0.5%
    533
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.16
    0.04
    533
    +0.10
    0.03
    425
    +0.10
    0.02
    Negative Logits
    iosyn
    -0.80
    QUI
    -0.75
    cise
    -0.74
    vic
    -0.72
    jong
    -0.71
    resy
    -0.71
    gu
    -0.69
    DAQ
    -0.68
    ourage
    -0.68
    must
    -0.67
    POSITIVE LOGITS
     obscurity
    1.03
     whence
    0.85
     afar
    0.84
     hating
    0.77
     humble
    0.75
     scratch
    0.70
     seed
    0.70
     pillar
    0.68
     conception
    0.68
     laughing
    0.68
    Activations Density 0.052%

    No Known Activations