© 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. 1611
    Prev
    Next
    INDEX
    Explanations

    phrases related to political campaigns

    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
    +0.14
    0.6%
    247
    +0.12
    0.5%
    526
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    247
    +0.14
    0.03
    371
    +0.12
    0.03
    723
    +0.11
    0.03
    Negative Logits
    galitarian
    -0.83
    ostic
    -0.82
    obyl
    -0.77
    ĪĴ
    -0.76
    ĨĴ
    -0.73
    ushima
    -0.73
    66666666
    -0.69
    orgetown
    -0.69
     Haram
    -0.68
     Tarant
    -0.66
    POSITIVE LOGITS
     slogan
    1.01
     trail
    1.00
     manager
    0.98
     finance
    0.97
     treasurer
    0.93
     campaigns
    0.86
     slogans
    0.86
     strategist
    0.83
    sters
    0.83
     rallies
    0.83
    Activations Density 0.040%

    No Known Activations