© 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. 24-GEMMASCOPE-2-RES-16K
    7. 8720
    Prev
    Next
    INDEX
    Explanations

    vulnerability prediction

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

    This neuron strongly activates on instructional or procedural language, particularly on “to” and nearby words in “how to”–style infinitive phrases.

    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
     каждый
    0.96
     только
    0.92
     dạng
    0.90
     стоимость
    0.90
    ოლოდ
    0.88
     tijekom
    0.88
     проверить
    0.88
    ၻ
    0.86
     каждым
    0.86
     позволяет
    0.85
    POSITIVE LOGITS
    ।
    0.81
    śmy
    0.81
    trim
    0.79
    west
    0.78
    yer
    0.76
    up
    0.75
    present
    0.74
    вав
    0.74
    𝗳
    0.72
    PID
    0.71
    Activations Density 0.000%

    No Known Activations