© 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. 10867
    Prev
    Next
    INDEX
    Explanations

    without explicit violence or harm

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

    The neuron is keyed to conditional or contingency wording—i.e. tokens that occur in “in case,” “if,” or similar clauses flagging potential problems, failures, or fallback scenarios.

    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
    ます
    0.71
    هههه
    0.67
    د
    0.61
    قا
    0.60
     resetCounters
    0.60
    יל
    0.59
    уга
    0.58
     seekBar
    0.57
    ﻥ
    0.56
    
    0.56
    POSITIVE LOGITS
     and
    0.95
     которые
    0.75
     or
    0.73
     alongside
    0.71
     culminating
    0.70
     as
    0.69
    scapes
    0.67
     большей
    0.66
     to
    0.65
     which
    0.65
    Activations Density 0.000%

    No Known Activations