© 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. 34-GEMMASCOPE-RES-131K
    7. 107103
    Prev
    Next
    INDEX
    Explanations

    they inform or act

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

    The neuron fires strongly on high-frequency function words—especially subject pronouns (“they,” “you,” “it”) and common auxiliaries/modals (“will,” “are,” “have,” “to,” “didn’t”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    paying
    -0.95
     gitara
    -0.94
     metav
    -0.91
     reisen
    -0.87
    lámpara
    -0.85
    ANDS
    -0.85
     Funktionen
    -0.84
     domanda
    -0.84
    ilige
    -0.83
    payers
    -0.82
    POSITIVE LOGITS
     send
    1.23
     sends
    1.20
     gave
    1.18
     give
    1.16
     records
    1.15
     tells
    1.15
     honoring
    1.14
     sending
    1.13
     honored
    1.12
     offer
    1.05
    Activations Density 0.030%

    No Known Activations