© 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. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 31-GEMMASCOPE-RES-16K
    6. 13453
    Prev
    Next
    INDEX
    Explanations

    references to numerical data or statistics

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

    numeric digits and DOI/URL reference identifiers in academic citations and technical documents.

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

    digits (particularly '7') within numeric identifiers, codes, or sequential numbers.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    numeric identifiers and codes in references/metadata (digits in URLs/DOIs, catalog or gene IDs, phone/patent numbers), with a strong bias toward the sequence 17.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    expandindo
    -0.56
    fromnode
    -0.48
     surla
    -0.45
    ustimmung
    -0.45
    addPreferredGap
    -0.40
    addContainerGap
    -0.40
    ⓧ
    -0.39
    __':
    
    -0.38
    новништво
    -0.37
    
    -0.36
    POSITIVE LOGITS
    rungsseite
    0.57
     chande
    0.48
     Taktlose
    0.44
    jooq
    0.43
     useRouter
    0.42
    IBOutlet
    0.41
    bodyParser
    0.41
     ContentValues
    0.40
    vider
    0.40
     Bec
    0.40
    Activations Density 0.187%

    No Known Activations