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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-27B-IT
    3. 31-GEMMASCOPE-2-RES-16K
    4. 11497
    Prev
    Next
    INDEX
    Explanations

    only always increases To

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

    The marked tokens appear across diverse document types (educational materials, exam questions, comprehension passages) and predominantly highlight parenthetical answer choices or their components, particularly closing parentheses, answer labels, and text within or immediately following answer options. The pattern reflects markup of multiple-choice question answer structure elements, especially focusing on option identifiers and their surrounding punctuation.

    eleuther_acts_top20 · claude-4-5-haikuTriggered by @jamesnaruto04

    multiple choice answer option text and content within answer choices.

    oai_token-act-pair · claude-4-5-haikuTriggered by @jamesnaruto04

    Incorrect answer options in multiple-choice questions within educational or assessment contexts, particularly those that contain factually wrong information, misleading statements, or mischaracterizations of concepts.

    eleuther_acts_top20 · claude-4-5-sonnetTriggered by @jamesnaruto04
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_31_width_16k_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
    というか
    0.52
     nutty
    0.50
     opsi
    0.47
    ก็
    0.47
     ו
    0.47
     също
    0.46
     også
    0.46
     också
    0.45
     poignant
    0.43
    টাও
    0.43
    POSITIVE LOGITS
     secreto
    0.42
    提高了
    0.42
     mejorar
    0.42
     aumentando
    0.41
     waardoor
    0.41
     wyłącznie
    0.38
    <unused2162>
    0.37
    keras
    0.37
    __
    0.37
     automatiquement
    0.37
    Activations Density 0.091%

    No Known Activations