© 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. 10-GEMMASCOPE-RES-131K
    4. 21744
    Prev
    Next
    INDEX
    Explanations

    instances and specific cases

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

    This neuron activates on abstract terms and modals that signal general requirements or occurrences in formal prose (e.g. necessity words like “must” or nouns denoting cases or presence).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     as
    -1.56
     three
    -1.48
     Hvis
    -1.42
     arba
    -1.42
     while
    -1.36
     parfü
    -1.34
     ترین
    -1.31
     what
    -1.30
     because
    -1.28
     menyadari
    -1.28
    POSITIVE LOGITS
     then
    2.17
     інших
    1.55
     zwykle
    1.47
    ród
    1.45
    となると
    1.44
    真的好
    1.41
     innych
    1.41
    withId
    1.36
     разных
    1.35
    我又
    1.34
    Activations Density 0.101%

    No Known Activations