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

    Neuronpedia

    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. 10015
    Prev
    Next
    INDEX
    Explanations

    ingredients and cooking instructions in a recipe

    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
    377
    +0.12
    0.6%
    187
    +0.12
    0.6%
    201
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.12
    0.01
    187
    +0.12
    0.01
    201
    +0.12
    0.01
    Negative Logits
    ensions
    -0.68
    ension
    -0.66
     prejudice
    -0.62
     Britann
    -0.62
     recourse
    -0.61
    fitting
    -0.60
     Cind
    -0.60
    ractive
    -0.59
     differential
    -0.57
    suit
    -0.57
    POSITIVE LOGITS
    igans
    1.01
    igan
    0.97
    ogly
    0.92
    cia
    0.92
    non
    0.91
    agher
    0.91
    za
    0.90
    culosis
    0.88
    ufact
    0.87
    istan
    0.85
    Activations Density 0.015%

    No Known Activations