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-2-9B-IT
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 13124
    Prev
    Next
    INDEX
    Explanations

    parts of sentences and formatting elements of text

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

    Bracketed expressions in code or text

    np_acts-logits-general · gemini-2.0-flash

    topics and concepts

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    Prompts (Dashboard)
    24,576 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
     Савезне
    -0.70
     Roskov
    -0.68
    AndEndTag
    -0.66
    IVEREF
    -0.58
     Winaray
    -0.57
    Décès
    -0.54
    HostException
    -0.50
    featureID
    -0.50
    StoreMessageInfo
    -0.50
     يتيمه
    -0.50
    POSITIVE LOGITS
    CreateModel
    0.46
     topic
    0.45
    Topic
    0.41
     Topic
    0.41
    AutoScale
    0.39
     durables
    0.39
    InMemory
    0.37
     profession
    0.35
    <0x86>
    0.35
     Mittag
    0.35
    Activations Density 0.765%

    No Known Activations