© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 35411
    Prev
    Next
    INDEX
    Explanations

    mathematical formulas and code variables

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

    The neuron chiefly recognizes numeric literals (numbers) in code.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    single-letter variables in mathematical expressions and code.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    single-letter variable identifiers in math or code expressions, especially when used as parameters in formulas or computations.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    5
    -1.07
     dziecka
    -1.02
     içi
    -1.02
    '
    -1.02
    々木
    -1.01
    8
    -1.00
    4
    -0.99
    niosek
    -0.99
    6
    -0.98
     ня
    -0.97
    POSITIVE LOGITS
    私たちは
    0.94
     satisfactor
    0.94
    тельным
    0.93
     solem
    0.93
    indeer
    0.90
    こういった
    0.89
    Син
    0.88
    見積
    0.88
     "").
    0.88
    !");
    0.88
    Activations Density 0.005%

    No Known Activations