© 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. 22-GEMMASCOPE-RES-131K
    4. 79521
    Prev
    Next
    INDEX
    Explanations

    police action and crime

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

    The neuron marks ordinary running‐text tokens in the body of a hard-news story—common words (names, verbs, pronouns, conjunctions, punctuation) found in narrative sentences—distinguishing them from non-article boilerplate or navigation text.

    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
     these
    -2.05
     that
    -1.94
     here
    -1.81
     we
    -1.64
     for
    -1.56
    Ethylbenzene
    -1.34
     artikke
    -1.34
     you
    -1.33
     today
    -1.33
     this
    -1.32
    POSITIVE LOGITS
     continuare
    1.41
     approx
    1.38
    lllllllll
    1.36
     animali
    1.34
    usions
    1.34
     two
    1.34
    llllllll
    1.33
    }});
    1.30
    vább
    1.28
    会计
    1.28
    Activations Density 0.018%

    No Known Activations