© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 20031
    Prev
    Next
    INDEX
    Explanations

    mentions of specific individuals and groups in relation to conspiracy theories and controversial viewpoints

    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
    639
    +0.10
    0.5%
    285
    +0.10
    0.4%
    440
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    525
    +0.10
    0.01
    353
    +0.10
    0.01
    639
    +0.10
    0.01
    Negative Logits
     Seym
    -0.82
    GROUND
    -0.76
    enance
    -0.70
    sburgh
    -0.69
    ij士
    -0.69
    lain
    -0.68
    ledged
    -0.67
    theless
    -0.66
     NETWORK
    -0.66
    EStream
    -0.65
    POSITIVE LOGITS
    otype
    1.12
    eman
    1.00
    emate
    0.97
    sey
    0.97
    ergic
    0.96
    ergy
    0.94
    iets
    0.94
    seed
    0.93
    nette
    0.91
    nea
    0.91
    Activations Density 0.021%

    No Known Activations