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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 31-GEMMASCOPE-2-RES-262K
    7. 15631
    Prev
    Next
    INDEX
    Explanations

    specific explanation

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

    tokens marking turn boundaries or speaker/role control tokens (e.g., <start_of_turn>, <end_of_turn>, "model", "user").

    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
     आफ्नो
    0.38
     hayat
    0.37
     सबै
    0.37
     ਆਪਣ
    0.37
     awhile
    0.36
     wares
    0.36
    విత
    0.36
    却没有
    0.36
     headache
    0.35
     आफ्
    0.35
    POSITIVE LOGITS
    Specific
    0.45
    具体
    0.45
     Specifically
    0.45
     denominada
    0.43
    specifically
    0.42
     предусматри
    0.41
     เนื่องจาก
    0.40
    specific
    0.40
    這種
    0.40
    具体的
    0.40
    Activations Density 0.217%

    No Known Activations