© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 12794
    Prev
    Next
    INDEX
    Explanations

    will + positive outcome verb

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

    The neuron is tuned to detect the modal auxiliary “will” (and closely associated promise-style verbs) used to make future-tense promises or assurances of benefit.

    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
     then
    -1.49
     it
    -1.30
     possibilities
    -1.16
     using
    -1.14
     instead
    -1.13
     vort
    -1.13
     possibilità
    -1.11
     possibility
    -1.09
    zonej
    -1.07
    lihood
    -1.05
    POSITIVE LOGITS
     you
    4.06
     your
    3.05
     вас
    1.79
     vás
    1.76
     вам
    1.72
     vám
    1.55
     YOU
    1.48
     आपको
    1.44
     vaše
    1.42
    あなたの
    1.36
    Activations Density 0.029%

    No Known Activations