© 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. 26184
    Prev
    Next
    INDEX
    Explanations

    themed locations or concepts

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

    The neuron strongly activates on content‐bearing topic words—single‐ or multiword nouns that label themes or categories (e.g. pirates, wild-west, transportation, community helpers, desert areas).

    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
     правительства
    -1.21
    fikation
    -1.13
    hangi
    -1.12
    })^
    -1.12
    Elke
    -1.06
     солнца
    -1.05
     américains
    -1.02
    ')")
    -1.02
    quelles
    -1.02
    自身の
    -1.02
    POSITIVE LOGITS
     themed
    1.63
     theme
    1.53
    themed
    1.30
     themes
    1.09
     Theme
    1.06
     темати
    1.05
     Have
    1.05
     weitere
    1.05
     motif
    1.04
     thème
    1.04
    Activations Density 0.046%

    No Known Activations