© 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. 394
    Prev
    Next
    INDEX
    Explanations

    start/end of phrases or sentences

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

    This neuron fires strongly on the very first meaningful content word of a news story or paragraph (i.e. the lead word in each segment).

    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
     recomand
    -0.90
     ,
    -0.88
     umum
    -0.85
    🛍
    -0.82
    🏷
    -0.82
    🤘
    -0.81
     nigger
    -0.80
    ₂,
    -0.80
    Algebra
    -0.79
    Displayed
    -0.79
    POSITIVE LOGITS
    </h4>
    1.16
    lognese
    0.96
     shouldn
    0.91
    啭
    0.91
    </h5>
    0.91
    你的
    0.90
    </tbody>
    0.90
     isn
    0.89
    (!__
    0.89
    </h1>
    0.88
    Activations Density 0.114%

    No Known Activations