© 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. 10-GEMMASCOPE-RES-131K
    7. 102472
    Prev
    Next
    INDEX
    Explanations

    indicates movement toward a goal

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

    The neuron selectively activates on directional “-ward” words—especially occurrences of “toward.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     
    -2.98
     m
    -2.81
     l
    -2.75
    ?"
    -2.72
     p
    -2.72
     t
    -2.55
     r
    -2.53
     time
    -2.50
     soo
    -2.50
     sly
    -2.45
    POSITIVE LOGITS
    鋨
    2.55
     \%)
    2.47
    阗
    2.39
     ſta
    2.33
     BARBARA
    2.20
    ляции
    2.20
     offrir
    2.20
     They
    2.19
    
    2.17
    étaient
    2.13
    Activations Density 0.008%

    No Known Activations