© 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 EvalsExportsinterp-engineNEWAPI 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. 2623
    Prev
    Next
    INDEX
    Explanations

    quotations from various sources

    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
    430
    +0.14
    0.6%
    121
    +0.10
    0.4%
    275
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    430
    +0.14
    0.05
    729
    +0.10
    0.04
    426
    +0.10
    0.04
    Negative Logits
    ingu
    -0.86
    etheless
    -0.83
    morph
    -0.74
    NetMessage
    -0.71
     sucker
    -0.71
    raph
    -0.70
    perature
    -0.70
     hindsight
    -0.70
    onom
    -0.70
     halluc
    -0.69
    POSITIVE LOGITS
     reads
    1.19
     wrote
    1.15
     tweeted
    0.95
     stated
    0.95
     said
    0.92
     commented
    0.87
     according
    0.87
    wrote
    0.86
     writes
    0.85
     spokeswoman
    0.83
    Activations Density 0.113%

    No Known Activations