Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-2B
    4. Residual Stream - 65k
    5. 20-GEMMASCOPE-RES-65K
    6. 778
    Prev
    Next
    INDEX
    Explanations

    terms and concepts related to quizzes and assessments

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    phrases and terms related to academic assessments and scores

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    exam and test preparation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.20.hook_resid_post
    Hook Layer
    20
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    LogFactory
    -0.47
    olymers
    -0.39
    AppendLine
    -0.39
    Sincerely
    -0.38
    owników
    -0.38
    lende
    -0.38
    chó
    -0.38
    )}_
    -0.37
     contado
    -0.37
    abag
    -0.35
    POSITIVE LOGITS
     exam
    2.18
     exams
    2.13
     examinations
    1.94
     examination
    1.93
     tests
    1.92
     Exam
    1.88
     Exams
    1.86
     test
    1.82
     Examination
    1.80
    exam
    1.79
    Activations Density 0.161%

    No Known Activations