© 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-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 11162
    Prev
    Next
    INDEX
    Explanations

    question marks and spanish endings

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

    This neuron lights up on unusually rare or low-frequency subword tokens—i.e. lexically uncommon pieces.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     conseg
    0.82
    蒉
    0.80
     nonce
    0.80
     réessayer
    0.80
     pratiquer
    0.79
    コラボ
    0.77
     próprios
    0.77
     invoc
    0.77
     tassels
    0.77
     gobl
    0.76
    POSITIVE LOGITS
    }$,
    1.00
    cs
    0.95
    also
    0.89
    cn
    0.77
    }$
    0.74
    cers
    0.73
    iseen
    0.73
    }$-
    0.73
    }$.
    0.71
    car
    0.70
    Activations Density 0.000%

    No Known Activations