© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 38599
    Prev
    Next
    INDEX
    Explanations

    leaving places or entities

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

    The neuron detects the direct object of “leave/left/leaving” constructions—that is, the thing or place being left.

    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
     ngào
    -1.11
     sorprender
    -1.10
     vêtements
    -1.09
    pasang
    -1.07
     demandes
    -1.06
    Against
    -1.06
     burbujas
    -1.05
     réve
    -1.04
     guste
    -1.03
    atting
    -1.02
    POSITIVE LOGITS
     to
    2.09
     into
    1.30
     during
    1.08
     after
    1.05
     via
    1.02
     by
    1.01
     shortly
    0.91
     faster
    0.87
     onto
    0.85
     uczniów
    0.84
    Activations Density 0.031%

    No Known Activations