© 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. 10-GEMMASCOPE-RES-131K
    4. 90871
    Prev
    Next
    INDEX
    Explanations

    ever, next, against

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

    The neuron strongly activates on subordinate or adverbial clauses—e.g. conditional “if…” phrases or prepositional modifiers (“from a mile away”) that add extra descriptive/detail context.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     cest
    -1.32
     to
    -1.27
     alcoh
    -1.25
     boya
    -1.25
    鱘
    -1.24
    錀
    -1.20
     roj
    -1.19
     rodi
    -1.16
    伱
    -1.16
     boty
    -1.15
    POSITIVE LOGITS
    就直接
    1.43
     them
    1.40
     prohibido
    1.28
    ($__
    1.27
    大于
    1.27
    ","+
    1.27
     rempliss
    1.26
    []"
    1.26
    ");
    1.25
    ');
    1.23
    Activations Density 0.012%

    No Known Activations