© 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. 3022
    Prev
    Next
    INDEX
    Explanations

    directions and locations

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

    The neuron detects references to location‐oriented instructions and headings—words like “Directions,” “Get,” and “address” that introduce or label navigational/placement information.

    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
    erende
    -0.73
    Deploy
    -0.73
    Tou
    -0.72
    ߜ
    -0.72
    eroon
    -0.71
     حس
    -0.71
    名は
    -0.71
    DOI
    -0.71
    ālā
    -0.71
     zey
    -0.70
    POSITIVE LOGITS
     directions
    5.16
     Directions
    4.16
    directions
    4.09
    Directions
    3.95
     DIRECTIONS
    3.11
     direction
    2.91
    DIRECTIONS
    2.89
     map
    2.64
    direction
    2.50
     maps
    2.47
    Activations Density 0.051%

    No Known Activations