© 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. 5992
    Prev
    Next
    INDEX
    Explanations

    information related to social and political issues, including healthcare, media coverage, and public relations

    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
    640
    +0.20
    0.9%
    373
    +0.11
    0.5%
    141
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.20
    0.08
    6
    +0.11
    0.05
    179
    +0.11
    0.04
    Negative Logits
    xual
    -0.74
     scrap
    -0.71
     orders
    -0.66
     custod
    -0.65
     attent
    -0.63
     honoured
    -0.63
     McMaster
    -0.61
     gard
    -0.61
     Maiden
    -0.61
     attentive
    -0.61
    POSITIVE LOGITS
    Sometimes
    1.11
    advertising
    1.05
    Unless
    1.04
    Firstly
    1.02
    Often
    0.99
    Reason
    0.99
    Probably
    0.99
    Everyone
    0.97
    Whenever
    0.97
    Again
    0.97
    Activations Density 0.189%

    No Known Activations