© 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 EvalsExportsinterp-engineNEWAPI 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. 16027
    Prev
    Next
    INDEX
    Explanations

    entities related to individuals or organizations in news or sports

    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
    447
    +0.34
    1.6%
    481
    +0.20
    1.0%
    691
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    481
    +0.34
    0.08
    288
    +0.20
    0.09
    146
    +0.11
    0.06
    Negative Logits
    keyes
    -0.56
    itton
    -0.55
    oya
    -0.55
    aughtered
    -0.54
    phies
    -0.52
    aughed
    -0.52
    uers
    -0.52
    urb
    -0.52
    hiba
    -0.51
    onde
    -0.51
    POSITIVE LOGITS
     lately
    0.80
    going
    0.66
     predicament
    0.65
     methodology
    0.64
    's
    0.63
     psyche
    0.63
     versus
    0.63
    Ê
    0.63
    esides
    0.59
     differ
    0.59
    Activations Density 0.643%

    No Known Activations