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. 2693
    Prev
    Next
    INDEX
    Explanations

    keywords related to programming functions and their definitions

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

    Code related to data validation and serialization

    np_acts-logits-general · gemini-2.0-flash

    programming verbs after punctuation

    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-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
    maal
    -0.55
     size
    -0.55
     target
    -0.55
     value
    -0.50
     turns
    -0.49
     total
    -0.49
    FontOfSize
    -0.49
     sc
    -0.49
    erequisite
    -0.48
     autor
    -0.47
    POSITIVE LOGITS
    Portale
    0.63
     Efq
    0.63
    ViewFeatures
    0.60
    Personensuche
    0.58
    LookAnd
    0.58
     Legendre
    0.58
     hdmi
    0.58
    فایل‌لار
    0.57
     gneiss
    0.57
     masts
    0.57
    Activations Density 0.148%

    No Known Activations