© 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. 5373
    Prev
    Next
    INDEX
    Explanations

    words related to dangerous or deadly situations or entities

    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
    719
    +0.13
    0.6%
    47
    +0.10
    0.5%
    174
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    719
    +0.13
    0.02
    343
    +0.10
    0.01
    32
    +0.09
    0.01
    Negative Logits
    ional
    -0.94
    ourced
    -0.89
    ational
    -0.87
    ibly
    -0.86
    bles
    -0.85
    ourcing
    -0.82
    edu
    -0.81
    é¾
    -0.76
    isse
    -0.75
    apest
    -0.74
    POSITIVE LOGITS
     whales
    0.92
    killer
    0.90
     instinct
    0.85
     spree
    0.84
     whale
    0.82
    knife
    0.79
     killer
    0.79
    fish
    0.76
     killers
    0.73
     blow
    0.72
    Activations Density 0.030%

    No Known Activations