© 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. 15606
    Prev
    Next
    INDEX
    Explanations

    sentence structures that begin with "It all comes together" or similar phrases indicating a cohesive analysis or summary

    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
    185
    +0.10
    0.4%
    641
    +0.09
    0.4%
    281
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    641
    +0.10
    0.02
    9
    +0.09
    0.02
    622
    +0.09
    0.02
    Negative Logits
    Torrent
    -0.76
    oiler
    -0.75
    \\\\
    -0.73
    rites
    -0.71
    ammy
    -0.70
    ëĭ
    -0.69
    ocument
    -0.69
    Interstitial
    -0.67
    emetery
    -0.67
    ufact
    -0.66
    POSITIVE LOGITS
     whether
    1.08
     motives
    1.06
     why
    0.97
     assumptions
    0.93
     legality
    0.87
    ibly
    0.86
    naires
    0.84
    whether
    0.82
     sincerity
    0.81
     validity
    0.81
    Activations Density 0.065%

    No Known Activations