© 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-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 95355
    Prev
    Next
    INDEX
    Explanations

    introducing specific concepts

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

    This neuron fires on the key descriptive word (noun or adjective) that immediately follows the phrase “with a.”

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

    the article "a" when it precedes a noun or noun phrase in formal or informational text.

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

    the indefinite article following “with,” i.e., the “with a/an …” construction.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    は
    -1.73
    这是一个
    -1.41
    が
    -1.40
     should
    -1.39
    焢
    -1.32
     with
    -1.31
    や
    -1.30
     オン
    -1.30
     dată
    -1.29
     there
    -1.28
    POSITIVE LOGITS
     regard
    2.03
     emphasis
    1.96
     regards
    1.75
    為に
    1.75
     impunity
    1.57
     respect
    1.55
    ering
    1.51
     help
    1.49
    ことが多い
    1.44
    outh
    1.43
    Activations Density 0.049%

    No Known Activations