© 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. 16503
    Prev
    Next
    INDEX
    Explanations

    phrases related to catastrophic events and significant problems

    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
    288
    +0.17
    0.8%
    447
    +0.12
    0.5%
    151
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.17
    0.02
    651
    +0.12
    0.02
    151
    +0.11
    0.02
    Negative Logits
    eton
    -0.84
    erd
    -0.83
    skirts
    -0.82
    ells
    -0.78
    ularity
    -0.78
    Ģ
    -0.77
    ular
    -0.77
    haw
    -0.75
    rab
    -0.74
    reenshots
    -0.73
    POSITIVE LOGITS
     catast
    1.06
     plag
    0.88
     engulf
    0.87
    ulence
    0.86
     inflicted
    0.85
     catastrophe
    0.84
     devast
    0.83
     relief
    0.83
     disasters
    0.83
     Outbreak
    0.82
    Activations Density 0.043%

    No Known Activations