© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 4263
    Prev
    Next
    INDEX
    Explanations

    words related to actions or events involving a significant amount of time

    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
    105
    +0.10
    0.4%
    17
    +0.09
    0.4%
    357
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    105
    +0.10
    0.03
    240
    +0.09
    0.03
    183
    +0.09
    0.03
    Negative Logits
    luaj
    -0.91
    tions
    -0.74
    holm
    -0.70
    è¦ļéĨĴ
    -0.70
    loop
    -0.70
    Americ
    -0.68
    lys
    -0.68
    ateg
    -0.68
    ibel
    -0.66
    ramid
    -0.65
    POSITIVE LOGITS
     awhile
    0.92
     precedence
    0.90
     longer
    0.84
     advantage
    0.82
     pains
    0.76
     aback
    0.75
    aways
    0.73
     time
    0.72
     heed
    0.72
     place
    0.70
    Activations Density 0.037%

    No Known Activations