© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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 3072
    9. 8-RES_FS3072-JB
    11. 720
    Prev
    Next
    INDEX
    Explanations

    mentions of public assistance and community support efforts

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    752
    +0.14
    0.6%
    698
    +0.11
    0.5%
    458
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    86
    +0.14
    0.09
    729
    +0.11
    0.08
    718
    +0.10
    0.11
    Negative Logits
     laughs
    -0.69
     perish
    -0.65
     airs
    -0.65
     Shutterstock
    -0.65
    ibles
    -0.62
     glances
    -0.60
     smiles
    -0.59
     lux
    -0.59
     afar
    -0.58
     lifeless
    -0.57
    POSITIVE LOGITS
     regarding
    1.14
     authorizing
    1.07
     concerning
    1.03
     recommending
    0.99
     approving
    0.96
    uggest
    0.95
     permitting
    0.95
     investigating
    0.91
     urging
    0.91
     identifying
    0.88
    Activations Density 3.886%

    No Known Activations