© 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. 36720
    Prev
    Next
    INDEX
    Explanations

    similarity, analogy, comparison

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

    The neuron fires on passages describing geometric “fit‐in” problems, i.e. mentions of shapes (like “cube”) and the phrasing around fitting one shape inside another.

    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
    others
    -0.89
     EventType
    -0.81
    ēs
    -0.77
     остальных
    -0.77
    xicity
    -0.77
     ulter
    -0.77
     supplémentaires
    -0.76
     SPORTS
    -0.75
     urat
    -0.75
     zwei
    -0.75
    POSITIVE LOGITS
     similar
    2.70
     similarly
    2.55
    similar
    2.28
     analogous
    2.23
     Similarly
    2.23
     Similar
    2.22
    Similar
    2.19
    Similarly
    2.16
     аналоги
    2.13
     podob
    2.05
    Activations Density 0.060%

    No Known Activations