© 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. 22503
    Prev
    Next
    INDEX
    Explanations

    references to Nazi ideology or groups

    oai_token-act-pair · gpt-4o-miniTriggered 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
    +0.16
    0.7%
    57
    +0.14
    0.6%
    481
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    57
    +0.16
    0.02
    288
    +0.14
    0.02
    146
    +0.13
    0.01
    Negative Logits
    tis
    -0.96
    lessly
    -0.92
    pring
    -0.78
    cially
    -0.78
    ascript
    -0.76
    above
    -0.75
    20439
    -0.74
    pole
    -0.73
    ately
    -0.72
    area
    -0.72
    POSITIVE LOGITS
     sympath
    0.93
     paramilitary
    0.91
     occupation
    0.89
     salute
    0.88
     ideology
    0.88
     swast
    0.84
     propaganda
    0.83
     atrocities
    0.81
     affili
    0.79
     uniforms
    0.78
    Activations Density 0.022%

    No Known Activations