© 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 24576
    9. 8-RES_FS24576-JB
    11. 19459
    Prev
    Next
    INDEX
    Explanations

    words related to controversial or negative information, particularly regarding political or societal issues

    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
    +0.11
    0.5%
    301
    +0.10
    0.5%
    481
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    151
    +0.11
    0.02
    532
    +0.10
    0.01
    655
    +0.10
    0.01
    Negative Logits
    arts
    -0.80
    ulton
    -0.79
     wisely
    -0.73
     Adviser
    -0.72
    ĺħ
    -0.72
    ĸļ
    -0.71
    agine
    -0.69
    gerald
    -0.68
    aido
    -0.68
    nan
    -0.68
    POSITIVE LOGITS
     hostility
    0.86
     disregard
    0.81
     racism
    0.78
     malice
    0.75
     refusal
    0.74
     contradiction
    0.73
     denial
    0.72
     rejection
    0.72
     ban
    0.71
     sexism
    0.71
    Activations Density 0.022%

    No Known Activations