© 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. 9779
    Prev
    Next
    INDEX
    Explanations

    questions or demands indicated by phrases starting with "Why" and "demand" or "insist"

    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
    665
    +0.10
    0.5%
    745
    +0.09
    0.4%
    737
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.10
    0.06
    49
    +0.09
    0.06
    718
    +0.09
    0.05
    Negative Logits
     Stru
    -0.76
     Laksh
    -0.70
     Rising
    -0.66
     Falling
    -0.65
    soDeliveryDate
    -0.64
     Lag
    -0.64
     Nationwide
    -0.63
     Eff
    -0.61
     Method
    -0.61
     Five
    -0.59
    POSITIVE LOGITS
     hear
    1.21
     see
    1.01
     borrow
    0.99
     emulate
    0.97
     know
    0.92
     speak
    0.89
     recreate
    0.89
     participate
    0.89
     abolish
    0.88
     receive
    0.88
    Activations Density 0.133%

    No Known Activations