© 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. 22-GEMMASCOPE-RES-131K
    7. 65560
    Prev
    Next
    INDEX
    Explanations

    Abnormal chemotaxis, Central Asia, Does matrix

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

    This neuron detects the numbers following the "+" in unified-diff hunk headers (i.e. the new-file line-count fields).

    oai_token-act-pair · o4-miniTriggered 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
    づけ
    -0.91
     Teraz
    -0.87
     of
    -0.87
     samfun
    -0.87
     planten
    -0.85
     март
    -0.85
     CONCLUSION
    -0.83
     geda
    -0.82
    ото
    -0.82
    collier
    -0.82
    POSITIVE LOGITS
    裄
    1.17
     recomand
    0.91
     аккуратно
    0.91
     pierdas
    0.88
    '],
    0.88
    acity
    0.87
     душу
    0.86
    実感
    0.84
    ikbaar
    0.84
    铿
    0.84
    Activations Density 0.000%

    No Known Activations