© 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. 5580
    Prev
    Next
    INDEX
    Explanations

    deal targeting younger fraudulent

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

    The neuron is a “content‐word” detector—it lights up strongly on semantically rich or uncommon tokens (verbs, adjectives, numbers, etc.) and stays off for common function words.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    phrases indicating someone is looking for or searching for something, particularly in the context of seeking information or solutions.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    terms indicating low amount, absence, or reduction (e.g., below/less/none states).

    oai_token-act-pair · gpt-5Triggered 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
     sclerosis
    0.76
     crap
    0.73
    ما
    0.71
     gem
    0.71
     snail
    0.71
     oyster
    0.70
     elderly
    0.70
     cliche
    0.69
     infantile
    0.68
     neuro
    0.68
    POSITIVE LOGITS
    ਤੀ
    0.92
    ﺪ
    0.89
    FI
    0.86
     fuese
    0.86
     подвер
    0.84
     ਜਾਣ
    0.84
     fueron
    0.83
     сразу
    0.83
     государственных
    0.82
    上がる
    0.82
    Activations Density 0.000%

    No Known Activations