© 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. 4522
    Prev
    Next
    INDEX
    Explanations

    network configuration and communication

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

    The neuron strongly activates on longer, formal technical or legal‐style abstract terms (e.g. “coherence,” “exaggerated,” “cloud-based,” “binding”), i.e. complex domain-specific 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
    e
    0.95
    ING
    0.89
    k
    0.89
    d
    0.88
    rb
    0.86
    r
    0.85
    se
    0.84
    ad
    0.83
    m
    0.83
    oin
    0.80
    POSITIVE LOGITS
     szolg
    0.79
    раў
    0.75
     radicales
    0.74
     socialism
    0.73
     peacetime
    0.73
     graphique
    0.73
     mitophagy
    0.72
    あなたは
    0.72
    ಸ್ತು
    0.71
     climatique
    0.71
    Activations Density 0.001%

    No Known Activations