© 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. 3279
    Prev
    Next
    INDEX
    Explanations

    references to names or titles of various individuals or entities

    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
    476
    +0.11
    0.5%
    63
    +0.11
    0.5%
    512
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    475
    +0.11
    0.06
    326
    +0.11
    0.06
    476
    +0.09
    0.04
    Negative Logits
    etheless
    -0.92
     guiName
    -0.77
     glim
    -0.64
    âĶĢâĶĢâĶĢâĶĢ
    -0.63
     tremend
    -0.62
    want
    -0.60
    âĸ¬âĸ¬
    -0.59
     tyr
    -0.59
     streng
    -0.58
    âķIJâķIJ
    -0.58
    POSITIVE LOGITS
    )
    2.07
    ),
    2.01
    )"
    1.97
    )'
    1.89
    ).
    1.87
    ),"
    1.85
    )."
    1.83
    )",
    1.83
    )—
    1.82
    )?
    1.78
    Activations Density 0.365%

    No Known Activations