© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 68913
    Prev
    Next
    INDEX
    Explanations

    harden, hardened, hardening

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

    The neuron activates on words and word parts related to “harden,” “hardening,” or “hardness.”

    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
    嵛
    -2.33
    ong
    -2.20
    </h3>
    -2.17
    bricas
    -2.11
    al
    -2.11
    electrical
    -2.03
    binant
    -1.99
     stát
    -1.98
     predetermined
    -1.97
    ᭝
    -1.95
    POSITIVE LOGITS
    );
    2.70
    </b>
    2.66
    。”
    2.39
    Ꭳ
    2.36
    ”;
    2.36
    !】
    2.27
    >
    
    2.20
    !”
    2.20
    羮
    2.16
    ︲
    2.16
    Activations Density 0.009%

    No Known Activations