© 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. 22254
    Prev
    Next
    INDEX
    Explanations

    quantities related to millions of people or dollars

    oai_token-act-pair · gpt-4o-miniTriggered 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
    71
    +0.12
    0.6%
    741
    +0.11
    0.5%
    218
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    741
    +0.12
    0.04
    71
    +0.11
    0.04
    98
    +0.10
    0.04
    Negative Logits
    Discuss
    -0.68
    krit
    -0.68
     distingu
    -0.67
     Dating
    -0.66
    imen
    -0.66
    ourage
    -0.65
    icut
    -0.64
    _>
    -0.64
    anwhile
    -0.64
    ainer
    -0.64
    POSITIVE LOGITS
     hectares
    1.16
     dollars
    1.15
     acres
    1.03
     tonnes
    1.02
     pounds
    1.01
    aires
    0.96
     gallons
    0.96
     lbs
    0.95
     sq
    0.92
     miles
    0.91
    Activations Density 0.060%

    No Known Activations