© 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. 10-GEMMASCOPE-RES-131K
    4. 1650
    Prev
    Next
    INDEX
    Explanations

    food preparation and processing

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

    food preparation actions like peeling, slicing, cleaning, and processing meat.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron fires on procedural or instructional keywords—especially verbs or section headings you’d see in recipes or other step‐by‐step guides.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    tiques
    -1.34
     MÉXICO
    -1.33
    怼
    -1.30
    ఞ
    -1.29
     appartement
    -1.29
    作战
    -1.24
    一丝
    -1.23
    不断的
    -1.21
    hline
    -1.20
     won
    -1.20
    POSITIVE LOGITS
    </h2>
    1.48
    }$.
    1.31
     and
    1.29
    已经
    1.26
    );
    1.26
    ')")
    1.24
    "].
    1.23
    閏
    1.23
    adesivo
    1.23
     speelt
    1.22
    Activations Density 0.035%

    No Known Activations