© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 64179
    Prev
    Next
    INDEX
    Explanations

    validation errors and code structure

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron activates on programming code tokens—things like import paths, class and function names, numbers, and other code identifiers—i.e. it’s picking out elements of source code.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    forgettable
    -0.83
    pkins
    -0.79
     cloak
    -0.76
     Khalid
    -0.74
     båda
    -0.73
     dialect
    -0.71
    的说
    -0.69
     Aes
    -0.69
    了口气
    -0.69
     Monmouth
    -0.69
    POSITIVE LOGITS
    gih
    0.86
     ער
    0.83
    aner
    0.81
    MaxLength
    0.80
    azah
    0.79
    erval
    0.78
     вку
    0.77
    Ensure
    0.76
     invi
    0.74
     besteht
    0.73
    Activations Density 0.005%

    No Known Activations