© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 79713
    Prev
    Next
    INDEX
    Explanations

    a specification using where

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

    The neuron detects relative‐clause markers—words like “whose,” “where,” and similar that introduce subordinate clauses.

    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
     conférence
    -1.09
     Described
    -1.07
     канала
    -1.05
     you
    -1.03
    🦤
    -1.02
     not
    -1.02
    bottomrule
    -1.00
     Instead
    -0.99
     rather
    -0.97
    スポンサーリンク
    -0.97
    POSITIVE LOGITS
     mỗi
    1.19
    each
    1.13
    每一
    1.11
     each
    1.10
    Each
    1.06
     каждой
    1.05
    !!!!!!!
    1.04
     chacun
    1.02
     각
    1.01
     marcó
    1.01
    Activations Density 0.050%

    No Known Activations