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

    Neuronpedia

    Jacobian LensNEW
    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. 43076
    Prev
    Next
    INDEX
    Explanations

    the word "whereas" in various contexts, indicating comparisons or contrasts

    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
    46
    +0.12
    0.5%
    662
    +0.09
    0.4%
    77
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    294
    +0.12
    0.01
    46
    +0.09
    0.01
    131
    +0.09
    0.01
    Negative Logits
    dyl
    -0.79
    anut
    -0.73
    packed
    -0.72
    hal
    -0.71
    atti
    -0.71
    acia
    -0.69
    atro
    -0.69
    eyed
    -0.69
    eat
    -0.68
    Synopsis
    -0.68
    POSITIVE LOGITS
    lihood
    0.87
     hitherto
    0.84
    xual
    0.81
    entimes
    0.81
     previous
    0.80
     previously
    0.80
     ordinarily
    0.77
     usual
    0.75
     Whereas
    0.74
     ours
    0.72
    Activations Density 0.013%

    No Known Activations