Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Gemma-3-1B-IT
    3. 17-GEMMASCOPE-2-RES-16K
    4. 160
    Prev
    Next
    INDEX
    Explanations

    instances of conversation/control tokens and turn boundaries (e.g., <start_of_turn>, system/model markers) — i.e., structural / meta tokens indicating speaker turns or formatting.

    oai_token-act-pair · gpt-5-miniTriggered by @toaster
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    gg-gs/gemma-scope-2-1b-it/resid_post
    Prompts (Dashboard)
    273,612 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     but
    1.17
     pero
    0.95
     maar
    0.89
     אבל
    0.88
    但我
    0.85
     yet
    0.84
    but
    0.84
     tetapi
    0.83
     }*/
    0.82
    이지만
    0.82
    POSITIVE LOGITS
    ↵↵
    2.06
    ↵
    1.17
    ↵↵↵
    1.04
    This
    0.83
    ↵↵↵↵
    0.70
    Several
    0.67
    Nearly
    0.66
    When
    0.65
     Virtually
    0.65
    Approximately
    0.64
    Activations Density 3.391%

    No Known Activations