© 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. 71456
    Prev
    Next
    INDEX
    Explanations

    heavier descriptors

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

    The neuron strongly responds to content‐heavy, multi-syllable words that carry descriptive or agentive meaning—especially those ending in “-or” (e.g. decorator), “-er” (e.g. heavier), or abstract skill nouns (e.g. dexterity, adept).

    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
    After
    -2.61
    Should
    -2.59
    -
    -2.56
            
    -2.55
    </b>
    -2.47
    ３
    -2.45
    -2.44
    0
    -2.42
    Since
    -2.42
    Of
    -2.42
    POSITIVE LOGITS
    髕
    2.66
     doesn
    2.61
     was
    2.59
    潆
    2.50
    s
    2.48
     isn
    2.47
     alot
    2.42
    该公司
    2.34
    氲
    2.30
     we
    2.27
    Activations Density 0.002%

    No Known Activations