© 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 EvalsExportsAPI 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. 10624
    Prev
    Next
    INDEX
    Explanations

    mentions of specific individuals or groups named "Sal" with various numbers attached

    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
    16
    +0.13
    0.6%
    314
    +0.11
    0.5%
    288
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    16
    +0.13
    0.02
    643
    +0.11
    0.02
    603
    +0.11
    0.02
    Negative Logits
     Dominion
    -0.83
    STEP
    -0.72
     depreciation
    -0.71
     clipboard
    -0.68
    lihood
    -0.68
     fixme
    -0.67
     bom
    -0.66
     GDDR
    -0.66
     Breaker
    -0.64
    hower
    -0.64
    POSITIVE LOGITS
    mone
    1.33
    omon
    1.32
    isbury
    1.26
    azar
    1.14
    adin
    1.13
    utations
    1.09
    ient
    1.06
    gado
    1.06
    amon
    1.05
    igon
    1.04
    Activations Density 0.010%

    No Known Activations