© 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. 9911
    Prev
    Next
    INDEX
    Explanations

    AI assistant responses

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

    This neuron detects tokens and phrases about AI systems, models, training/data, and system-level instructions or outputs (e.g., "model", "assistant", "input", "data", "generated", "training", "server", "json").

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31
    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
    றுத்த
    0.38
    কর্ণ
    0.36
    আ
    0.36
    しまった
    0.36
    屣
    0.36
    勍
    0.36
    ্ব
    0.35
    ाइवेट
    0.34
    ର୍
    0.34
     éviter
    0.34
    POSITIVE LOGITS
    ،
    0.48
    ፣
    0.40
     hingegen
    0.40
     byly
    0.39
    частли
    0.38
    мые
    0.38
     italiano
    0.38
    that
    0.38
    тику
    0.38
     ,
    0.37
    Activations Density 0.373%

    No Known Activations