© 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. 73874
    Prev
    Next
    INDEX
    Explanations

    involve a, to invasive, tablets running, phosphatase activity

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

    mathematical notations and scientific content in academic or technical documents.

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

    The neuron is keyed to spotting the letter sequence “ion” (i.e. nominalizations ending in –ion) inside 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
     inconce
    -1.06
    徊
    -1.02
    既然
    -1.02
    Auszeichnungen
    -0.96
     horrend
    -0.94
     imperceptible
    -0.93
     Чтобы
    -0.93
     incompar
    -0.92
     Hence
    -0.92
     Its
    -0.91
    POSITIVE LOGITS
     different
    1.15
    回来了
    1.12
     cliquant
    1.02
     ecles
    1.02
     various
    1.01
     highlighted
    1.00
     motocicletas
    0.99
     on
    0.99
    -),
    0.98
     this
    0.98
    Activations Density 0.052%

    No Known Activations