© 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. 13982
    Prev
    Next
    INDEX
    Explanations

    numbers, titles, and Q&A

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

    The neuron is keying on hyphens (– or -) used as separators in numbered or lettered headings and list markers (e.g. “1 - …”, “A1 - …”, “– …”).

    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
     your
    -1.95
     éte
    -1.84
     you
    -1.84
     @"/
    -1.80
    </u>
    -1.72
    esinde
    -1.70
    jenige
    -1.70
    
    -1.63
     it
    -1.61
    いえば
    -1.59
    POSITIVE LOGITS
     "
    1.74
     Eventually
    1.74
     
    1.74
     observes
    1.70
    gered
    1.64
     但
    1.56
     enlarged
    1.48
     examines
    1.47
     $
    1.43
    </h2>
    1.42
    Activations Density 0.000%

    No Known Activations