© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 14761
    Prev
    Next
    INDEX
    Explanations

    One Minute, DAY

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

    The neuron fires on the short, punchy standalone words or phrases (often capitalized) used as dramatic taglines—e.g. the one‐word or two‐word emphatic “sentences” in trailers like “Looks. Intense.”

    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
     kompatibel
    -1.18
     anhänger
    -1.16
     recevrez
    -1.11
     allerlei
    -1.11
     vokal
    -1.10
     autres
    -1.09
     semblent
    -1.08
     pourront
    -1.08
     sekunder
    -1.08
    jds
    -1.07
    POSITIVE LOGITS
     decisiones
    1.09
     tiny
    1.06
     incredible
    1.04
     people
    1.03
    })$
    1.02
    univ
    1.01
     пишет
    1.01
     häufig
    1.00
     conductas
    1.00
     As
    0.99
    Activations Density 0.017%

    No Known Activations