© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 5424
    Prev
    Next
    INDEX
    Explanations

    about or towards + identifying entities

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

    The neuron strongly activates on numeric expressions (especially multi‐digit numbers and decimals), i.e. figures and statistics.

    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
    Ві
    0.72
     কী
    0.67
    ~/
    0.66
    엔
    0.66
    ज़
    0.66
    וא
    0.66
    П
    0.66
    但
    0.64
    直播
    0.63
    ắ
    0.63
    POSITIVE LOGITS
     evidences
    1.32
     equipments
    1.28
     stuffs
    1.12
     researches
    1.09
     aforesaid
    1.08
     loosing
    1.08
     advices
    1.08
     sufferings
    1.05
     softwares
    1.03
     appare
    0.99
    Activations Density 1.304%

    No Known Activations