© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 40-GEMMASCOPE-2-RES-262K
    7. 12195
    Prev
    Next
    INDEX
    Explanations

    instructions and rules for systems

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

    instructions that try to override the model's rules or assign it a persona (jailbreak/prompt‑injection commands telling the model to ignore safety constraints).

    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_40_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
     വിദ
    0.49
     เนื่อง
    0.48
    нік
    0.46
     мен
    0.45
    িং
    0.45
     Astrophys
    0.43
     بُ
    0.42
     speak
    0.42
     케
    0.42
    ниците
    0.42
    POSITIVE LOGITS
     alerts
    0.41
     ready
    0.41
    Ready
    0.41
     rounds
    0.40
    Amt
    0.39
    ready
    0.38
    amt
    0.38
    운
    0.38
     specials
    0.38
    苓
    0.38
    Activations Density 25.365%

    No Known Activations