© 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. 28448
    Prev
    Next
    INDEX
    Explanations

    last word in a phrase

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

    The neuron fires strongly on the head noun of a noun phrase—i.e. on the key noun that names the category or entity in a phrase.

    oai_token-act-pair · o4-miniTriggered 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
    ).
    -1.73
    .”
    -1.57
     etc
    -1.54
    时长
    -1.23
    </i>
    -1.22
    传统的
    -1.22
    ”.
    -1.21
     to
    -1.19
    ".
    -1.19
    </h2>
    -1.18
    POSITIVE LOGITS
     Publicado
    1.41
     azonban
    1.40
    Ꮚ
    1.37
    beur
    1.36
     nhưng
    1.34
    かもしれませんが
    1.32
     lutar
    1.31
     mantan
    1.28
     pö
    1.28
     but
    1.27
    Activations Density 0.231%

    No Known Activations