© 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. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 34999
    Prev
    Next
    INDEX
    Explanations

    physical cell identifier

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

    The neuron fires on content‐bearing technical terms—especially scientific or mathematical nouns (e.g. “muscle,” “atrophy,” “spin,” “network,” “rotation,” “intensity,” etc.).

    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
    poptosis
    -1.02
     ``
    -0.97
    这场
    -0.89
    Ќ
    -0.87
     назвать
    -0.85
     intensiv
    -0.82
     "
    -0.82
    本日の
    -0.80
     unreachable
    -0.80
     `
    -0.79
    POSITIVE LOGITS
     relaciona
    1.08
     różnych
    1.02
     интере
    0.98
    ↼
    0.96
    ajajaja
    0.93
    もあり
    0.93
    odorante
    0.92
    澆
    0.91
     colocar
    0.90
     colocando
    0.89
    Activations Density 0.083%

    No Known Activations