© 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. 2994
    Prev
    Next
    INDEX
    Explanations

    positive and negative appraisals

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

    The neuron is detecting open‐class content words (nouns, verbs, adjectives, etc.) as opposed to closed‐class or function words.

    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
     agus
    0.95
     illetve
    0.88
     utilises
    0.87
     pith
    0.83
     banjo
    0.83
     (
    0.82
     alot
    0.82
     bloke
    0.82
    6
    0.82
     odnosno
    0.81
    POSITIVE LOGITS
    恐怕
    0.86
    💑
    0.83
    かもしれません
    0.80
    這樣的
    0.79
    🙆
    0.79
     sufferings
    0.79
    こんな
    0.78
    🕌
    0.77
    Вот
    0.75
     pitiful
    0.75
    Activations Density 0.585%

    No Known Activations