© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 75266
    Prev
    Next
    INDEX
    Explanations

    and + noun

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

    This neuron selectively activates on multi‐syllabic abstract nouns—which are often nominalizations of verbs (e.g. planning, distribution, building, function).

    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
     when
    -1.66
     if
    -1.53
     cuando
    -1.37
    fordshire
    -1.28
     både
    -1.27
     before
    -1.20
     they
    -1.20
    ).
    -1.17
     but
    -1.16
     because
    -1.12
    POSITIVE LOGITS
    んですよね
    1.40
     pensare
    1.30
    んですね
    1.27
    んでしょう
    1.27
    んだろう
    1.24
    ようです
    1.24
    んですよ
    1.23
    わけです
    1.18
    _
    1.14
     actions
    1.13
    Activations Density 0.107%

    No Known Activations