© 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 12288
    9. 8-RES_FS12288-JB
    11. 10710
    Prev
    Next
    INDEX
    Explanations

    negative commentary related to political figures

    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
    117
    +0.12
    0.6%
    631
    +0.09
    0.5%
    682
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    117
    +0.12
    0.05
    374
    +0.09
    0.03
    682
    +0.09
    0.04
    Negative Logits
     largeDownload
    -0.77
    viation
    -0.71
    pmwiki
    -0.67
    ruary
    -0.65
    etheless
    -0.65
     Franç
    -0.64
    theless
    -0.63
    bara
    -0.61
    elin
    -0.61
    âĹ¼
    -0.61
    POSITIVE LOGITS
     Kavanaugh
    1.00
     Dems
    0.80
    Ħ¢
    0.71
    ']
    0.70
    ').
    0.69
    %]
    0.68
     accuser
    0.68
     Replay
    0.67
     probe
    0.66
     Rosenstein
    0.64
    Activations Density 0.118%

    No Known Activations