© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 25469
    Prev
    Next
    INDEX
    Explanations

    mathematical constraints and derivations

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

    The neuron is sensitive to numeric tokens and related mathematical labels (e.g. numbers, numeric literals, “constraint,” “following,” and colons) in equation contexts.

    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
     fédé
    -1.13
     controladores
    -1.10
    Recept
    -1.02
    licas
    -1.00
     exactement
    -0.94
     deswegen
    -0.93
    Termine
    -0.92
     всички
    -0.91
    िक्ष
    -0.91
     persönlich
    -0.90
    POSITIVE LOGITS
     also
    1.34
     before
    1.24
     from
    1.16
     additional
    1.11
     as
    1.02
     Also
    1.02
     or
    1.00
     Before
    0.99
     by
    0.94
     relationship
    0.94
    Activations Density 0.018%

    No Known Activations