© 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. 127947
    Prev
    Next
    INDEX
    Explanations

    This neuron responds to content‐bearing words (open‐class tokens like nouns, verbs, and adjectives) and stays off for common function words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    but or who

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     this
    -1.02
     when
    -0.98
    Zubereitung
    -0.98
     chanteuse
    -0.94
     tetap
    -0.93
     Sénat
    -0.92
    verty
    -0.92
     sempre
    -0.91
    −/−
    -0.89
    uose
    -0.88
    POSITIVE LOGITS
     known
    1.35
     who
    1.09
     neither
    1.06
    神经
    1.05
    臜
    1.02
     personalidade
    1.02
     old
    1.00
    neither
    0.97
     ugyan
    0.96
     reemplazo
    0.96
    Activations Density 0.028%

    No Known Activations