© 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 12288
    9. 8-RES_FS12288-JB
    11. 9882
    Prev
    Next
    INDEX
    Explanations

    references to political relationships and international cooperation

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    393
    +0.13
    0.6%
    362
    +0.11
    0.5%
    375
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.13
    0.06
    59
    +0.11
    0.05
    611
    +0.09
    0.05
    Negative Logits
    fuck
    -0.71
    ucci
    -0.66
    boxing
    -0.65
    Joined
    -0.64
    forge
    -0.64
    webkit
    -0.63
     deserves
    -0.62
     istg
    -0.61
    igue
    -0.61
    .''.
    -0.60
    POSITIVE LOGITS
     hostilities
    0.80
     these
    0.80
     certain
    0.78
     those
    0.69
     the
    0.67
    interstitial
    0.66
     such
    0.65
     sorts
    0.64
    ints
    0.62
     goods
    0.61
    Activations Density 0.304%

    No Known Activations