© 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. 15337
    Prev
    Next
    INDEX
    Explanations

    mentions of payment, compensation, and benefits in relation to employees, especially federal workers

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    424
    +0.12
    0.6%
    218
    +0.10
    0.5%
    70
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    424
    +0.12
    0.02
    70
    +0.10
    0.02
    648
    +0.09
    0.02
    Negative Logits
    GGGGGGGG
    -0.69
     Uriel
    -0.67
     Veg
    -0.66
    âĸ¬
    -0.64
     Nare
    -0.63
    Untitled
    -0.61
     Frankfurt
    -0.61
    selves
    -0.61
     Twist
    -0.61
     Beir
    -0.59
    POSITIVE LOGITS
    checks
    1.37
    outs
    1.26
    offs
    1.02
    roll
    1.00
    paying
    0.97
    Scale
    0.94
    back
    0.94
    paid
    0.91
    backs
    0.90
     earners
    0.89
    Activations Density 0.025%

    No Known Activations