© 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. 13364
    Prev
    Next
    INDEX
    Explanations

    adverbs that emphasize or intensify the meaning of the following words

    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
    363
    +0.12
    0.5%
    106
    +0.09
    0.4%
    68
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.12
    0.05
    106
    +0.09
    0.04
    333
    +0.09
    0.03
    Negative Logits
    usterity
    -0.80
    igger
    -0.80
    utenberg
    -0.79
    yne
    -0.72
    ibble
    -0.71
    umo
    -0.69
    actionDate
    -0.67
    raught
    -0.66
    ocalyptic
    -0.65
    achev
    -0.65
    POSITIVE LOGITS
     invests
    1.14
     expects
    1.14
     enjoys
    1.09
     proceeded
    1.08
     thereafter
    1.05
     regulates
    1.04
     intends
    1.04
     employs
    1.03
     avoids
    1.03
     replaces
    1.02
    Activations Density 0.163%

    No Known Activations