© 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. 11160
    Prev
    Next
    INDEX
    Explanations

    qualifying adverbs and modal verbs

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

    The neuron fires on words that convey speaker attitude or emphasis—subjective qualifiers and intensifiers like “really,” “feel,” “want,” “need,” “actually,” “terrible,” “of course,” etc.

    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
     such
    -1.58
     into
    -1.55
     of
    -1.55
    :
    -1.50
     to
    -1.47
    ;
    -1.35
    !
    -1.33
     here
    -1.30
     around
    -1.30
    ).
    -1.27
    POSITIVE LOGITS
     amerikan
    1.40
    »).
    1.37
     Singapur
    1.32
     entran
    1.30
     gæ
    1.27
    emplares
    1.26
    ﻤ
    1.26
    sehr
    1.26
    ÇÃO
    1.26
    »;
    1.22
    Activations Density 0.081%

    No Known Activations