© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 16425
    Prev
    Next
    INDEX
    Explanations

    headings and labels

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

    This neuron fires on prominent nouns that denote destinations or goals, especially in directional or guiding phrases (e.g. “in any game,” “to the crossroads,” “to the way,” “to the Lord’s Table,” “to these new platforms”).

    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
     there
    -1.18
     pensamentos
    -0.92
    來自
    -0.92
     chegar
    -0.90
    ㄟ
    -0.87
     vergeten
    -0.87
     formalized
    -0.87
     polecam
    -0.86
    整體
    -0.85
    katakan
    -0.85
    POSITIVE LOGITS
     ihrer
    1.14
     by
    1.11
     their
    1.09
    abilidade
    1.05
    使其
    1.04
    ghouse
    1.02
     tanpa
    1.02
     mereka
    1.02
    esian
    0.99
     via
    0.98
    Activations Density 0.092%

    No Known Activations