© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 53-GEMMASCOPE-2-RES-262K
    7. 50448
    Prev
    Next
    INDEX
    Explanations

    bomb making instructions

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

    sentences where the model declares safety constraints/refuses to comply (e.g., "I am programmed to be a safe and helpful AI assistant" and similar refusal/policy statements).

    oai_token-act-pair · gpt-5-mini
    Triggered by @yooniel31
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_53_width_262k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    Gap
    0.40
    罡
    0.39
     ইয়াহিয়া
    0.39
    uştur
    0.38
     ステ
    0.38
    gap
    0.37
     الأرض
    0.37
    Noah
    0.35
     Gap
    0.35
    рек
    0.35
    POSITIVE LOGITS
     tass
    0.41
     really
    0.40
     monatomic
    0.40
     GIR
    0.38
    প্রস
    0.38
     obj
    0.37
     Ransom
    0.37
     crash
    0.36
     Ride
    0.36
     ransomware
    0.36
    Activations Density 0.003%

    No Known Activations