© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 14938
    Prev
    Next
    INDEX
    Explanations

    phrases related to significant dates and events, specifically New Year's Eve

    oai_token-act-pair · gpt-3.5-turbo
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    575
    +0.12
    0.5%
    206
    +0.11
    0.5%
    393
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    664
    +0.12
    0.02
    575
    +0.11
    0.02
    206
    +0.10
    0.01
    Negative Logits
    nels
    -0.84
    STER
    -0.78
    selves
    -0.74
    inka
    -0.70
    aja
    -0.69
    ACP
    -0.68
    INST
    -0.67
    argon
    -0.66
    FILE
    -0.65
    NER
    -0.64
    POSITIVE LOGITS
     sake
    0.91
     Eve
    0.89
     worth
    0.87
    pace
    0.84
    peed
    0.69
    cue
    0.67
     finest
    0.65
     Day
    0.64
    piration
    0.63
     anniversary
    0.63
    Activations Density 0.047%

    No Known Activations