© 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. 93309
    Prev
    Next
    INDEX
    Explanations

    article titles and descriptions

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

    The neuron fires on the opening words of a new section or headline—i.e. tokens at the start of titles or section headers.

    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
     luks
    -1.41
     klien
    -1.27
     seksi
    -1.22
     ][
    -1.21
    tonsoft
    -1.21
    Режи
    -1.20
    宜しくお願い
    -1.16
     beinahe
    -1.16
     հղ
    -1.15
     ekos
    -1.15
    POSITIVE LOGITS
    </b>
    1.52
    "
    1.29
     -
    1.23
    <b>
    1.21
    В
    1.20
     с
    1.19
    ан
    1.18
    с
    1.14
    —
    1.12
     will
    1.10
    Activations Density 0.060%

    No Known Activations