© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 78947
    Prev
    Next
    INDEX
    Explanations

    citation closing punctuation

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

    The neuron fires on long numeric tokens—specifically the multi-digit reference or identifier numbers (e.g. citation IDs or PMIDs) embedded in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
     многих
    -1.18
    色々と
    -1.10
     vários
    -1.10
     sejumlah
    -1.00
     verschillende
    -0.99
    
    -0.99
    haden
    -0.97
     wielu
    -0.96
     nombreuses
    -0.95
     numerosos
    -0.95
    POSITIVE LOGITS
    )
    1.00
    \},
    0.94
     stared
    0.94
    ],
    0.93
     who
    0.93
    矗
    0.92
     quinta
    0.90
    لينكات
    0.89
    talet
    0.88
    ),
    0.88
    Activations Density 0.027%

    No Known Activations