© 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. 22-GEMMASCOPE-RES-131K
    7. 72817
    Prev
    Next
    INDEX
    Explanations

    relative pronouns and conjunctions

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

    This neuron fires strongly on the very first content word of a sentence (typically the sentence’s subject).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    relative clauses introduced by "that" followed by a verb, especially in formal or technical writing.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    references to prior entities via relative/subordinate clauses and pronoun+auxiliary constructions (e.g., “that/which/where” clauses or “you/they + are/have”).

    oai_token-act-pair · gpt-5Triggered 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
    </em>
    -1.15
    </strong>
    -1.08
     вся
    -1.02
     debido
    -0.96
     to
    -0.92
     also
    -0.91
                         
    -0.90
     throughout
    -0.90
    olkata
    -0.88
     उन
    -0.87
    POSITIVE LOGITS
    明日は
    1.14
    畕
    1.13
    mado
    1.13
    "+
    1.13
    ánk
    1.13
    Gdy
    1.09
    昨日は
    1.09
    Está
    1.09
    にほんブログ村
    1.07
    吅
    1.07
    Activations Density 0.073%

    No Known Activations