© 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. 67712
    Prev
    Next
    INDEX
    Explanations

    a [word] that

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

    The neuron fires on the main noun that begins an answer—i.e. the subject word immediately following the “A” marker in Q&A responses.

    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
     what
    -1.68
     most
    -1.39
     any
    -1.37
     udang
    -1.23
     kupa
    -1.21
     much
    -1.18
     anyone
    -1.18
    gunakan
    -1.18
     bulu
    -1.17
     obviamente
    -1.16
    POSITIVE LOGITS
    ซึ่ง
    1.25
    p
    1.09
     ainsi
    1.09
     et
    1.09
     campagnes
    1.09
    了一眼
    1.07
     товары
    1.03
    UNIFIED
    1.03
     oth
    1.02
    li
    1.01
    Activations Density 0.434%

    No Known Activations