© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 107677
    Prev
    Next
    INDEX
    Explanations

    pronouns followed by to

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

    The neuron flags the key tokens (like “you,” “him,” “to,” and “want”) in constructions of the form “X wants Y to …,” i.e. expressions of desire with a target and an infinitival “to.”

    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
     willingness
    -1.03
     érde
    -0.97
    予定
    -0.95
     Bereits
    -0.95
     defensor
    -0.94
     previously
    -0.94
    让我
    -0.94
    está
    -0.94
     sellado
    -0.93
     zostały
    -0.91
    POSITIVE LOGITS
     to
    2.38
     badly
    1.15
     done
    1.05
     involved
    1.05
     gone
    1.01
     velkommen
    0.99
     alive
    0.98
     delivered
    0.91
     replaced
    0.91
    radiated
    0.90
    Activations Density 0.044%

    No Known Activations