© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 15052
    Prev
    Next
    INDEX
    Explanations

    models trained

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

    This neuron fires most strongly on words occurring right at the start of a new paragraph or text segment, with its activation decaying the farther you get into the segment.

    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
    izens
    0.85
    ตก
    0.84
    𝚕
    0.81
    část
    0.80
     ফুলে
    0.80
    ើត
    0.77
     Fools
    0.77
    𝚘
    0.76
    াপাশি
    0.75
    junctive
    0.75
    POSITIVE LOGITS
    TOR
    0.79
    ان
    0.77
     TRI
    0.77
    RBI
    0.76
    積極的に
    0.75
    TRI
    0.74
     oameni
    0.74
    RIP
    0.72
     CLIENTI
    0.72
    OD
    0.71
    Activations Density 0.000%

    No Known Activations