© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 1708
    Prev
    Next
    INDEX
    Explanations

    except but instead

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

    The neuron fires on contrast/exception markers and adjoining qualifiers (words like “except,” “but,” “worse,” “plus,” “just”) that signal a shift or contrast in discourse.

    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
     mesmo
    -1.25
     similar
    -1.18
    asiest
    -1.10
     même
    -1.08
     despite
    -0.99
     traditional
    -0.99
    酊
    -0.97
    同様
    -0.95
    /$',
    -0.94
     strong
    -0.94
    POSITIVE LOGITS
     instead
    3.03
     вместо
    2.45
    Instead
    2.19
     Instead
    2.03
     with
    2.02
     zamiast
    1.96
    instead
    1.84
     istället
    1.75
     invece
    1.62
     bardziej
    1.57
    Activations Density 0.075%

    No Known Activations