© 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. 43086
    Prev
    Next
    INDEX
    Explanations

    references to the term "overnight."

    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
    288
    +0.14
    0.7%
    481
    +0.11
    0.5%
    651
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    651
    +0.14
    0.01
    178
    +0.11
    0.01
    158
    +0.10
    0.01
    Negative Logits
    umbn
    -0.82
    dinand
    -0.82
    sembly
    -0.81
    ktop
    -0.77
    boa
    -0.77
    orius
    -0.75
    artisan
    -0.75
    iov
    -0.73
    ioxide
    -0.73
    rera
    -0.73
    POSITIVE LOGITS
    aneously
    1.00
     overnight
    0.79
     oats
    0.78
    time
    0.78
     commute
    0.74
     turnaround
    0.72
     shift
    0.70
    noon
    0.70
     ow
    0.70
     raids
    0.69
    Activations Density 0.005%

    No Known Activations