© 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. 22-GEMMASCOPE-RES-131K
    4. 70704
    Prev
    Next
    INDEX
    Explanations

    contrasting conjunctions or affirmations

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

    The neuron fires on content words inside concessive or comparative subordinate clauses (e.g. “although that is certainly a side benefit” or “as much as that should peeve the Cowboys”).

    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
     not
    -1.23
     only
    -1.16
     Unlike
    -1.14
     unlike
    -1.09
     from
    -1.02
     with
    -0.96
     Since
    -0.93
     either
    -0.91
     before
    -0.90
     Despite
    -0.88
    POSITIVE LOGITS
    確かに
    1.26
     nhưng
    1.12
     but
    1.10
     certamente
    1.09
    Certainly
    0.98
    thats
    0.98
    BUT
    0.93
     แต่
    0.93
    それで
    0.93
    もありますが
    0.91
    Activations Density 0.040%

    No Known Activations