© 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. 67545
    Prev
    Next
    INDEX
    Explanations

    trivial statements or affirmations

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

    The neuron fires on downplaying qualifiers—especially “just” (and similar minimizers like “only” or “another”) used to diminish or dismiss a following noun phrase.

    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
     gracilis
    -0.76
    头像
    -0.74
    UIManager
    -0.73
     Bres
    -0.73
    fillRect
    -0.72
    ちゃんと
    -0.72
    孢
    -0.71
     guineas
    -0.71
    確
    -0.70
    ricts
    -0.69
    POSITIVE LOGITS
    又不是
    1.00
     merely
    0.94
    ظهار
    0.94
    品です
    0.91
    Although
    0.90
    tował
    0.85
    Betyg
    0.85
    不过是
    0.84
    Neden
    0.83
    trivial
    0.83
    Activations Density 0.039%

    No Known Activations