© 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. 7141
    Prev
    Next
    INDEX
    Explanations

    I am digging a hole

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

    sentences where the assistant refuses a harmful request or provides safety/warning language (e.g., "I am programmed to be a safe/harmless AI" and related resources).

    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
    pyridine
    0.19
    »),
    0.18
     glycer
    0.18
    δ
    0.18
     derajat
    0.17
     حقیقت
    0.17
    }),
    0.17
    त्मक
    0.17
    “,
    0.17
    =',
    0.17
    POSITIVE LOGITS
     welkom
    0.19
     कॅ
    0.18
    苏
    0.18
    মন্দ
    0.17
     그냥
    0.17
    ↵↵↵
    0.17
    पेट
    0.17
     колдон
    0.17
    ール
    0.17
    トイレ
    0.17
    Activations Density 0.582%

    No Known Activations