© 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 49152
    9. 8-RES_FS49152-JB
    11. 41606
    Prev
    Next
    INDEX
    Explanations

    references to specific dish or culinary terms

    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
    289
    +0.10
    0.5%
    17
    +0.10
    0.5%
    286
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    289
    +0.10
    0.00
    658
    +0.10
    0.01
    190
    +0.10
    0.01
    Negative Logits
    istically
    -0.81
     ital
    -0.69
     Mayhem
    -0.66
     Archdemon
    -0.65
    */(
    -0.65
     Extended
    -0.64
     Destination
    -0.63
    ciating
    -0.62
    akeru
    -0.62
    isations
    -0.61
    POSITIVE LOGITS
    washer
    1.60
    onest
    1.35
    enger
    1.08
    wall
    1.04
    ennes
    0.94
    ware
    0.91
    washing
    0.89
    arm
    0.88
    renheit
    0.87
    OME
    0.86
    Activations Density 0.004%

    No Known Activations