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 - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13527
    Prev
    Next
    INDEX
    Explanations

    active verbs related to outdoor and indoor activities, particularly emphasizing nighttime contexts and potential dangers

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    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
    qua
    -0.49
    곧
    -0.45
    cshtml
    -0.43
    eradish
    -0.42
    acke
    -0.41
    っていきます
    -0.41
    istory
    -0.41
    ثار
    -0.41
    디어
    -0.40
    resiasi
    -0.40
    POSITIVE LOGITS
     daytime
    2.07
     nighttime
    1.95
     evenings
    1.87
     night
    1.81
     evening
    1.75
     mornings
    1.68
     nights
    1.63
     afternoons
    1.62
     nachts
    1.62
     nocturnal
    1.56
    Activations Density 0.403%

    No Known Activations