© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 68706
    Prev
    Next
    INDEX
    Explanations

    in situ, vivo, vitro, accessible

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

    scientific or technical terms that are used in vitro, in vivo, or in reference to digital environments.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron flags uncommon, content‐heavy tokens—typically longer, domain-specific words rather than frequent function 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
     étanche
    -0.84
    arcas
    -0.81
    itettura
    -0.79
     Abigail
    -0.78
    Bö
    -0.77
     ਆ
    -0.76
    ТВ
    -0.76
    rese
    -0.76
    Terminate
    -0.75
     enclose
    -0.74
    POSITIVE LOGITS
     accessible
    1.05
     located
    0.92
    zuka
    0.80
     méthode
    0.78
    腭
    0.75
    ύν
    0.75
     accessed
    0.75
    فته
    0.75
    撰文
    0.73
    duma
    0.72
    Activations Density 0.106%

    No Known Activations