© 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. 5716
    Prev
    Next
    INDEX
    Explanations

    The neuron appears to detect computations involving division and floating-point numbers, often denoted by 'f' or 'F'

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

    The main thing this neuron does is pick out numeric tokens—particularly decimal or floating-point numbers.

    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
     whirl
    0.52
     outburst
    0.52
     zile
    0.48
    处
    0.47
    ंटा
    0.47
    姑娘
    0.47
    τώ
    0.46
     whirlwind
    0.46
     ragazze
    0.46
     voie
    0.46
    POSITIVE LOGITS
    ⣿
    0.73
    ⣶
    0.61
    angsa
    0.54
    鞆
    0.54
    ப்படம்
    0.52
    𝑭
    0.52
    ensing
    0.51
    
    0.51
     f
    0.50
    культур
    0.50
    Activations Density 0.532%

    No Known Activations