© 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. 16214
    Prev
    Next
    INDEX
    Explanations

    references to specific years

    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
    640
    +0.16
    0.8%
    447
    +0.14
    0.6%
    183
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    183
    +0.16
    0.06
    123
    +0.14
    0.06
    257
    +0.11
    0.06
    Negative Logits
     lodged
    -0.69
     endors
    -0.68
     nailed
    -0.67
     unemploy
    -0.66
     behav
    -0.65
     outwe
    -0.65
     combatants
    -0.63
     Rohing
    -0.63
     entertained
    -0.63
     distur
    -0.63
    POSITIVE LOGITS
    ventory
    1.45
    cluding
    1.40
    strument
    1.29
    visible
    1.25
    cluded
    1.25
    clude
    1.25
    juries
    1.24
    vasive
    1.24
    sect
    1.22
    iti
    1.21
    Activations Density 0.127%

    No Known Activations