© 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. 24-GEMMASCOPE-2-RES-16K
    7. 1072
    Prev
    Next
    INDEX
    Explanations

    commands and lists

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

    The neuron is highly sensitive to common function words and pronouns used in direct, promotional or persuasive address (e.g. “you,” “will,” “keep,” “attention,” “our,” “the”), effectively flagging reader-directed marketing language.

    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
    اا
    0.50
    𝙾
    0.50
    utilizzo
    0.49
     ทีม
    0.49
     всё
    0.48
    ₍
    0.47
     व्हाउचर
    0.46
     סט
    0.46
     รวมถึง
    0.46
     Ա
    0.46
    POSITIVE LOGITS
     psychotherapy
    0.35
     antihy
    0.34
     physiologically
    0.33
     nutritive
    0.33
     arterios
    0.33
     atherosclerosis
    0.32
     prolet
    0.32
     systolic
    0.31
     volatil
    0.31
     constitutive
    0.31
    Activations Density 0.004%

    No Known Activations