© 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. 23998
    Prev
    Next
    INDEX
    Explanations

    terms and phrases related to animosity, discrimination, and violence, particularly towards marginalized 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
    679
    +0.10
    0.4%
    516
    +0.09
    0.4%
    736
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.10
    0.06
    225
    +0.09
    0.05
    679
    +0.09
    0.06
    Negative Logits
     curv
    -0.76
     sidx
    -0.73
    amera
    -0.69
    isher
    -0.68
     ellipt
    -0.68
    amen
    -0.64
     tentative
    -0.63
     weight
    -0.62
    BuyableInstoreAndOnline
    -0.61
     nutrit
    -0.61
    POSITIVE LOGITS
     militias
    0.77
    hirt
    0.77
     emanating
    0.77
     abroad
    0.76
     sects
    0.74
     extremism
    0.70
    eering
    0.70
    ivism
    0.69
    Ń·
    0.68
     alike
    0.68
    Activations Density 0.284%

    No Known Activations