© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 16332
    Prev
    Next
    INDEX
    Explanations

    foreign words and phrases

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

    The neuron fires on the central “action” word or its direct object in how-to or technical sentences—i.e. the main verb or the noun being acted on (chicken, DNA, patient, program, dog, shoot, etc.).

    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
     and
    -1.67
    為に
    -1.16
     it
    -1.14
    を目指す
    -1.04
    conocimiento
    -1.03
     excepto
    -1.01
     燒
    -1.00
     suuri
    -0.98
     všetko
    -0.98
    ようで
    -0.98
    POSITIVE LOGITS
     into
    1.28
    を
    1.18
    ですが
    1.13
    などは
    1.09
    が
    1.09
    などを
    1.01
     плата
    0.99
    は
    0.97
    があります
    0.96
     promessa
    0.96
    Activations Density 0.083%

    No Known Activations