© 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. 130618
    Prev
    Next
    INDEX
    Explanations

    The neuron selectively activates on title-case segments—proper names and key words in titles of works, programs, or formal acts.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    acts and legislation

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     decorativo
    -1.42
    voici
    -1.38
     enthou
    -1.27
    givenchy
    -1.25
     Kälte
    -1.20
     plumas
    -1.20
     ungewöhn
    -1.20
    酲
    -1.20
     perles
    -1.17
    besos
    -1.17
    POSITIVE LOGITS
    ^{}
    1.54
    new
    1.30
    ierzch
    1.21
     More
    1.16
    <tr>
    1.16
    riezmann
    1.16
    ucleotide
    1.16
     —
    1.15
    ticos
    1.15
    ><
    1.14
    Activations Density 0.078%

    No Known Activations