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

    Neuronpedia

    Jacobian LensNEW
    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. 17803
    Prev
    Next
    INDEX
    Explanations

    proper nouns related to politics, particularly the term "Obama"

    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
    362
    +0.13
    0.6%
    539
    +0.10
    0.4%
    598
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    343
    +0.13
    0.03
    247
    +0.10
    0.03
    472
    +0.09
    0.03
    Negative Logits
    NetMessage
    -0.97
    terday
    -0.71
    requires
    -0.70
    duty
    -0.68
    perse
    -0.67
    ...]
    -0.66
    agree
    -0.66
     Ranked
    -0.66
    abwe
    -0.65
     Takeru
    -0.65
    POSITIVE LOGITS
     administration
    1.26
     Administration
    1.07
     era
    1.06
     presidency
    1.04
     campaign
    1.01
     playbook
    1.01
     agenda
    0.98
     dynasty
    0.93
     regime
    0.92
     administrations
    0.91
    Activations Density 0.090%

    No Known Activations