© 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. 34-GEMMASCOPE-RES-131K
    4. 82519
    Prev
    Next
    INDEX
    Explanations

    material states and properties

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

    The neuron fires on technical adjectives and nouns describing material states or properties (e.g. “fibrous,” “crystalline,” “carbonaceous,” “amorphous,” “gaseous,” “liquid,” “solid,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    ≈
    -0.99
     Idee
    -0.94
     idee
    -0.88
     واح
    -0.87
    ogues
    -0.84
     arrivent
    -0.84
     dma
    -0.84
     arriva
    -0.83
    ource
    -0.82
    gruppe
    -0.82
    POSITIVE LOGITS
     material
    1.20
     whatnot
    1.11
     chcete
    1.10
     artesanía
    1.08
     creatures
    1.07
     materials
    1.06
    事物
    1.00
     sustancias
    0.99
    ingue
    0.98
     automóviles
    0.97
    Activations Density 0.207%

    No Known Activations