© 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. 67412
    Prev
    Next
    INDEX
    Explanations

    exceptions to positivity

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

    This neuron responds strongly to emphasis and formatting markers—unusually capitalized words, colons, parentheses, and other punctuation used to highlight or set off 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
     veste
    -1.04
    boxylic
    -1.01
    ⛪
    -0.96
    系的
    -0.94
    vės
    -0.92
     วิ
    -0.91
    ּוֹ
    -0.91
     보
    -0.91
     arit
    -0.91
    !]
    -0.90
    POSITIVE LOGITS
     except
    1.80
    except
    1.70
     aunque
    1.45
    although
    1.38
     but
    1.37
     Except
    1.37
     although
    1.29
     allerdings
    1.29
     only
    1.27
     burbujas
    1.25
    Activations Density 0.098%

    No Known Activations