© 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. 14725
    Prev
    Next
    INDEX
    Explanations

    adjectives related to intelligence or trickery

    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
    99
    +0.10
    0.5%
    447
    +0.10
    0.4%
    723
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    292
    +0.10
    0.03
    499
    +0.10
    0.02
    151
    +0.09
    0.03
    Negative Logits
    imester
    -0.69
    asta
    -0.69
    IGH
    -0.64
    ategory
    -0.63
    ICAN
    -0.63
    thur
    -0.63
    thood
    -0.63
     Alive
    -0.62
    astered
    -0.62
    aru
    -0.61
    POSITIVE LOGITS
    ly
    0.98
     tricks
    0.91
     ploy
    0.91
     tact
    0.89
    NESS
    0.87
     twist
    0.87
     manipulation
    0.86
     satire
    0.86
    nesses
    0.85
     twists
    0.85
    Activations Density 0.082%

    No Known Activations