© 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. 3196
    Prev
    Next
    INDEX
    Explanations

    phrases related to health insurance and medical programs, particularly mentioning the organization "Blue Cross Blue Shield."

    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
    407
    +0.10
    0.5%
    72
    +0.08
    0.4%
    510
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.10
    0.05
    103
    +0.08
    0.04
    482
    +0.08
    0.04
    Negative Logits
    cens
    -0.70
     CY
    -0.63
    aceae
    -0.60
     Reaction
    -0.58
    vernment
    -0.58
    abilia
    -0.56
     IPM
    -0.56
    ;}
    -0.55
     AUD
    -0.54
    izable
    -0.53
    POSITIVE LOGITS
     knees
    0.67
    arde
    0.66
    akers
    0.62
     shoulders
    0.58
     riches
    0.57
    aker
    0.57
     bones
    0.57
     chest
    0.57
     arrow
    0.56
    FT
    0.56
    Activations Density 0.566%

    No Known Activations