© 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 49152
    9. 8-RES_FS49152-JB
    11. 42905
    Prev
    Next
    INDEX
    Explanations

    specific references to geopolitical themes and issues

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    480
    +0.45
    2.2%
    288
    +0.12
    0.6%
    384
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.45
    0.09
    480
    +0.12
    0.01
    266
    +0.10
    0.06
    Negative Logits
    姫
    -0.63
     Leone
    -0.62
    DonaldTrump
    -0.62
    terday
    -0.61
    è£ħ
    -0.60
    iannopoulos
    -0.57
     Cosponsors
    -0.57
    ccording
    -0.57
     Rica
    -0.56
    lda
    -0.54
    POSITIVE LOGITS
     itself
    1.08
    iest
    0.92
     thereof
    0.91
     portion
    0.82
    liest
    0.82
     aspect
    0.82
     side
    0.79
     ones
    0.78
    osphere
    0.77
     realm
    0.76
    Activations Density 0.460%

    No Known Activations