© 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. 10897
    Prev
    Next
    INDEX
    Explanations

    linguistic variation

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

    The neuron lights up on relatively uncommon content words (longer, low‐frequency terms).

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

    commas that separate clauses or items in complex, formal written text.

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

    function words and punctuation that introduce or structure clauses—like relative/subordinate or interrogative markers, and parenthetical separators (dashes/quotes).

    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
    !
    0.92
    ه
    0.84
    i
    0.79
     gangster
    0.75
     mafia
    0.75
    லூர்
    0.74
     despot
    0.72
     fracking
    0.72
     billionaire
    0.71
     telep
    0.71
    POSITIVE LOGITS
     Specifies
    0.81
    一直是
    0.79
    ভৃতি
    0.75
    範囲
    0.74
     আরম্ভ
    0.74
    క్కువ
    0.73
    滗
    0.71
    明显的
    0.71
     包括
    0.71
    是在
    0.71
    Activations Density 0.287%

    No Known Activations