© 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. 34-GEMMASCOPE-RES-131K
    7. 31176
    Prev
    Next
    INDEX
    Explanations

    compression and encoding

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

    The neuron is detecting code tokens—programming identifiers, keywords, and symbols—rather than ordinary prose.

    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
    חר
    -0.78
     spout
    -0.77
    UITextField
    -0.73
    맘
    -0.73
    åringen
    -0.73
    сування
    -0.72
    כז
    -0.71
    Constitu
    -0.71
    Mauricio
    -0.71
    RAND
    -0.70
    POSITIVE LOGITS
     compressed
    1.47
    compressed
    1.31
    压缩
    1.13
     Compressed
    1.12
     run
    1.06
     encoded
    1.02
     Compression
    1.02
    compression
    1.02
     encoding
    1.01
     compression
    0.99
    Activations Density 0.034%

    No Known Activations