© 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. 9522
    Prev
    Next
    INDEX
    Explanations

    threatening note

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

    This neuron selectively lights up on narrative-style prose—especially the opening words of descriptive sentences in a story (character names or pronouns, articles like “The” or “A,” and key action/setting words).

    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
     also
    -1.80
     its
    -1.34
    also
    -1.33
    </h2>
    -1.15
    Also
    -1.14
     to
    -1.13
     também
    -1.12
     and
    -1.09
     también
    -1.09
     where
    -1.06
    POSITIVE LOGITS
    一旦
    1.05
    hahahaha
    1.01
     sembler
    1.00
    再度
    0.96
     abomination
    0.95
     MUCH
    0.94
    uksen
    0.94
     Factbook
    0.94
     kilomètres
    0.94
    nością
    0.93
    Activations Density 0.003%

    No Known Activations