© 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. 120
    Prev
    Next
    INDEX
    Explanations

    corrections or inaccuracies in news stories

    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
    723
    +0.11
    0.5%
    28
    +0.10
    0.5%
    267
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    723
    +0.11
    0.02
    306
    +0.10
    0.02
    267
    +0.09
    0.02
    Negative Logits
    soDeliveryDate
    -0.81
    Mods
    -0.69
     cabinets
    -0.64
    Applications
    -0.64
    mods
    -0.64
     candy
    -0.63
     anyways
    -0.63
    ateurs
    -0.62
     suites
    -0.61
     subconscious
    -0.60
    POSITIVE LOGITS
     corrected
    0.91
     typo
    0.91
     incorrectly
    0.86
     incorrect
    0.84
    orrect
    0.83
    Clar
    0.79
     inaccur
    0.78
     spelling
    0.78
     cler
    0.77
     omitted
    0.75
    Activations Density 0.096%

    No Known Activations