© 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. 12700
    Prev
    Next
    INDEX
    Explanations

    except as otherwise when

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

    The neuron fires on words signaling exclusion or negation—e.g. “except,” “unless,” “not,” “without”—marking exception clauses or negative qualifiers.

    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
    dienne
    -1.03
    ounces
    -0.99
    très
    -0.97
     vôtre
    -0.96
    licitations
    -0.96
     doivent
    -0.95
    ľov
    -0.94
    oplasmic
    -0.94
     malheureusement
    -0.94
    OnAdd
    -0.93
    POSITIVE LOGITS
     prescribed
    1.16
     previously
    1.15
     if
    1.13
     black
    1.10
     ausdrücklich
    1.08
     可以
    1.03
     according
    1.02
     accepted
    1.02
     aquellos
    1.01
     traditionally
    1.00
    Activations Density 0.051%

    No Known Activations