© 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. 39341
    Prev
    Next
    INDEX
    Explanations

    could say

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

    The neuron fires on hedge or stance‐marking verbs and phrases (e.g. “could,” “argue,” “say,” “think”) that signal speculation or opinion.

    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
    e
    -1.73
    ↵↵↵↵↵↵↵↵↵↵↵↵
    -1.63
    {
    -1.51
    ↵↵↵↵↵↵↵↵↵
    -1.50
    ↵↵↵↵↵↵
    -1.48
    for
    -1.45
    ↵↵↵↵↵↵↵
    -1.42
    ↵↵↵
    -1.42
    G
    -1.41
    ↵↵↵↵↵↵↵↵↵↵↵
    -1.38
    POSITIVE LOGITS
     also
    1.66
     retrô
    1.58
     prenez
    1.55
     óculos
    1.50
     protège
    1.45
     klap
    1.39
     akumulator
    1.38
     veneno
    1.37
     garantit
    1.34
     rús
    1.34
    Activations Density 0.015%

    No Known Activations