© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 38820
    Prev
    Next
    INDEX
    Explanations

    hardness, stiffness, tartness

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

    The neuron fires on terms that describe material properties—especially hardness, stiffness/rigidity, or other technical descriptors often accompanied by numeric measures.

    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
    ">+
    -0.98
    közi
    -0.93
     Δι
    -0.92
     esquer
    -0.92
    Friday
    -0.90
     colorida
    -0.89
    zew
    -0.88
     aliviar
    -0.88
    sighs
    -0.88
    uebas
    -0.88
    POSITIVE LOGITS
     sharp
    1.53
     harsh
    1.48
     sharper
    1.42
     sharpening
    1.32
     sharpness
    1.32
     rigid
    1.17
     sharpened
    1.16
    Sharper
    1.16
    rigid
    1.14
    sharp
    1.13
    Activations Density 0.135%

    No Known Activations