© 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. 98379
    Prev
    Next
    INDEX
    Explanations

    greetings and punctuation

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

    The neuron fires on informal discourse markers or interjections (e.g. “yes,” “no,” “right,” “man,” “shall”) typically appearing at the start of an utterance.

    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
     despite
    -1.20
    最近は
    -1.12
    uchtung
    -1.11
     since
    -1.10
    '}),
    -1.05
    assment
    -1.04
    ActionPerformed
    -1.04
    Eine
    -1.03
    هران
    -1.02
    Erstellt
    -1.01
    POSITIVE LOGITS
    .
    1.33
    pecially
    1.20
     mensen
    1.18
     retira
    1.17
     meines
    1.16
    But
    1.15
     puramente
    1.09
     caminar
    1.09
    了下去
    1.09
    muñeca
    1.09
    Activations Density 0.001%

    No Known Activations