© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 5146
    Prev
    Next
    INDEX
    Explanations

    autotroph, fine-tuning, point

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

    The neuron detects the initial segments of hyphenated compound words (e.g. the left‐half before “-in,” “-off,” “-retardant,” “-fiction,” etc.).

    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
     Bonds
    0.46
    ﺔ
    0.41
     Gemeinde
    0.41
     Debate
    0.41
     Formen
    0.40
     Keeper
    0.40
    affiliated
    0.40
     Département
    0.40
     Initiate
    0.40
     qtd
    0.40
    POSITIVE LOGITS
    ኬ
    0.38
     thể
    0.38
    ırı
    0.38
    。”
    0.37
    hitungan
    0.36
    。"
    0.35
     massively
    0.34
     пола
    0.34
     khoản
    0.33
     tùy
    0.33
    Activations Density 0.000%

    No Known Activations