© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B-IT
    7. Residual Stream - 131k
    9. 20-GEMMASCOPE-RES-131K
    11. 50107
    Prev
    Next
    INDEX
    Explanations

    key terms and phrases related to technology and its applications

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

    anthropological state

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron activates on content-bearing words — especially nouns, technical terms, proper nouns, and other informative lexical tokens (as opposed to function words).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    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
    adaptiveStyles
    -0.62
    ContentAsync
    -0.59
    WebElementEntity
    -0.56
     للاسماء
    -0.56
     мәкал
    -0.54
    AccessorTable
    -0.53
     CallOverrides
    -0.53
     nahilalakip
    -0.53
     gynhyrchwyd
    -0.52
     يتيمه
    -0.50
    POSITIVE LOGITS
     Hoffnung
    0.38
    הערות
    0.37
     yaşında
    0.34
     voyageurs
    0.34
    PathVariable
    0.34
     Medicina
    0.34
     onCreateView
    0.33
     Bischof
    0.33
     prowadzi
    0.32
    InputTagHelper
    0.32
    Activations Density 0.138%

    No Known Activations