© 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. 10941
    Prev
    Next
    INDEX
    Explanations

    specific terms

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

    The neuron lights up on high‐salience, summary‐style or announcement words—i.e. the key action‐oriented and topic-signaling tokens (like “Advice:”, “approved”, “plan”, “forward”, “overall outlook”, “promising”, etc.) that mark main points or section headings.

    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
    τζ
    0.84
     chodzi
    0.82
     चूंकि
    0.81
    दरअसल
    0.80
    不对
    0.80
    是怎么
    0.78
    RIK
    0.77
    𝗜
    0.75
    𝗟
    0.75
    𝐏
    0.75
    POSITIVE LOGITS
    ៕
    0.89
     semoga
    0.88
     conformance
    0.88
     continue
    0.81
     môže
    0.78
     Xer
    0.75
     cements
    0.74
    <eos>
    0.73
     fournir
    0.73
     podrá
    0.72
    Activations Density 7.648%

    No Known Activations