© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 561
    Prev
    Next
    INDEX
    Explanations

    contrast or opposition

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

    The neuron flags common “function” words—small grammatical tokens like pronouns (I, my), conjunctions (and), modals/auxiliaries (will, might, has), and adverbs (just, along).

    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
     only
    -1.16
     keeps
    -0.95
     so
    -0.93
     brings
    -0.91
    Carcinogenicity
    -0.90
    Mutagenicity
    -0.89
     ALWAYS
    -0.86
     ensures
    -0.86
     just
    -0.84
    Urls
    -0.84
    POSITIVE LOGITS
     แต่
    1.46
     даже
    1.45
     nhưng
    1.44
     nawet
    1.44
     but
    1.42
     zelfs
    1.28
     certamente
    1.25
     навіть
    1.24
     incluso
    1.24
     لكن
    1.23
    Activations Density 0.042%

    No Known Activations