© 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. 72553
    Prev
    Next
    INDEX
    Explanations

    cardinal directions and grid positions

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

    The neuron mainly fires on words indicating directional motion (e.g. “move,” “up,” “right,” “down”).

    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
    archical
    -1.55
    erobic
    -1.42
    Technologie
    -1.27
    Aplica
    -1.20
    Garantie
    -1.20
     tecnológica
    -1.19
    Característica
    -1.17
    wendungs
    -1.17
    نية
    -1.16
    érape
    -1.13
    POSITIVE LOGITS
     left
    1.52
     right
    1.34
    gencias
    1.23
    licante
    1.20
     opposite
    1.19
    він
    1.12
     east
    1.11
     вперед
    1.10
    nemia
    1.09
     ἐκ
    1.09
    Activations Density 0.014%

    No Known Activations