© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 31-GEMMASCOPE-2-RES-262K
    7. 4793
    Prev
    Next
    INDEX
    Explanations

    words following certain tokens

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

    It detects long or rare subword tokens—often capitalized, technical, or proper‑name fragments (e.g., domain-specific/uppercase chunks).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_31_width_262k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     to
    0.25
     for
    0.24
     on
    0.18
    to
    0.18
    r
    0.17
    1
    0.17
    is
    0.16
    র
    0.16
     Japan
    0.15
     at
    0.15
    POSITIVE LOGITS
    的的
    0.17
    рьох
    0.17
     muñ
    0.16
    ჲ
    0.16
    Ꭸ
    0.16
     beak
    0.16
    йної
    0.16
    neho
    0.16
    ॓
    0.16
     maxillary
    0.15
    Activations Density 5.862%

    No Known Activations