© 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. 374
    Prev
    Next
    INDEX
    Explanations

    which followed by verbs

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

    The neuron is detecting the relative pronoun “which” (i.e. instances introducing 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
     then
    -1.29
    では
    -1.18
     it
    -1.16
    ++.
    -1.16
     protože
    -1.15
     sudadera
    -1.13
    eamiento
    -1.11
    ureusement
    -1.11
     būti
    -1.09
     cápsulas
    -1.08
    POSITIVE LOGITS
    ようです
    1.30
    んですね
    1.21
    !
    1.12
    ようになりました
    1.09
    のでしょう
    1.07
     You
    1.05
    のですね
    1.04
     Some
    1.04
    ↵↵↵
    0.93
     originally
    0.91
    Activations Density 0.052%

    No Known Activations