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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 103
    Prev
    Next
    INDEX
    Explanations

    asking questions

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

    The neuron primarily fires on conjugated verb forms—i.e. verbs inflected with endings like “-ed” or “-s.”

    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
     државе
    -0.93
    along
    -0.93
    ことから
    -0.89
     terceros
    -0.88
     próprio
    -0.88
     خودم
    -0.86
    alker
    -0.85
    ster
    -0.85
    of
    -0.85
     indeb
    -0.84
    POSITIVE LOGITS
     his
    1.80
     their
    1.65
     its
    1.50
     atau
    1.16
     seus
    1.13
    📲
    1.08
     himself
    1.07
     нам
    1.05
     our
    1.05
    brücken
    1.05
    Activations Density 0.075%

    No Known Activations