© 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. 68378
    Prev
    Next
    INDEX
    Explanations

    i don't think

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

    This neuron detects first‐person opinion or hedging phrases, especially “I think” (and its variants like “I don’t think”).

    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
     bereits
    -1.44
    鎶
    -1.41
    縝
    -1.39
    牴
    -1.38
     становника
    -1.36
    expédition
    -1.36
    谘
    -1.36
     adverten
    -1.35
     trochu
    -1.35
    暱
    -1.34
    POSITIVE LOGITS
     it
    2.03
     they
    1.80
     you
    1.49
    2
    1.47
    //!
    
    1.34
    —
    1.34
     anyone
    1.31
     anc
    1.31
    ,
    1.26
    ­
    1.26
    Activations Density 0.010%

    No Known Activations