© 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. 9122
    Prev
    Next
    INDEX
    Explanations

    adjectives and phrases related to social or political issues and controversies

    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
    87
    +0.34
    1.8%
    222
    +0.10
    0.6%
    331
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.34
    -0.02
    331
    +0.10
    0.05
    222
    +0.09
    0.03
    Negative Logits
     IMAGES
    -0.57
    eters
    -0.57
    é¾į
    -0.54
    Downloadha
    -0.54
     subsequently
    -0.52
    hran
    -0.52
    displayText
    -0.52
    )|
    -0.48
    imedia
    -0.48
    -0.48
    POSITIVE LOGITS
    izational
    0.61
    ueless
    0.58
     coincidence
    0.57
     iceberg
    0.57
    establishment
    0.57
     paradise
    0.57
     underdog
    0.55
    iterranean
    0.55
     dystop
    0.55
     minus
    0.54
    Activations Density 0.682%

    No Known Activations