© 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. 14577
    Prev
    Next
    INDEX
    Explanations

    concepts and processes

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

    The neuron fires on content-carrying words (i.e. non–function words) such as nouns and verbs that carry the main semantic load rather than on common grammatical tokens.

    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
     sportifs
    -1.14
     transparencia
    -1.13
     egenskaper
    -1.10
     colgar
    -1.10
    años
    -1.09
     verksamhet
    -1.07
    郇
    -1.06
     réserver
    -1.05
     там
    -1.05
    zusetzen
    -1.05
    POSITIVE LOGITS
     to
    1.05
    rightarrow
    1.03
    ️⃣
    0.99
     और
    0.98
     Secrétaire
    0.97
     ت
    0.96
    T
    0.95
     urging
    0.94
     advantage
    0.94
     other
    0.94
    Activations Density 0.018%

    No Known Activations