© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 10407
    Prev
    Next
    INDEX
    Explanations

    7 followed by rescue numbers

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

    The neuron is specially tuned to detect standalone numeric tokens—particularly numerical measurements (e.g., frequency values, ranges, IDs) in the text.

    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_31_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
     jonka
    0.68
     pct
    0.68
     भाग
    0.67
    瘦
    0.63
    體
    0.62
     نگه
    0.61
     Astro
    0.60
     الرغم
    0.59
    攔
    0.59
    affen
    0.58
    POSITIVE LOGITS
    concentration
    0.66
    MAPE
    0.66
    СО
    0.64
    НК
    0.64
    Establishing
    0.64
    Establish
    0.63
    Perspective
    0.62
     Ге
    0.61
    Restrictions
    0.61
    MAE
    0.61
    Activations Density 0.086%

    No Known Activations