© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    3. Gemma-2-9B
    4. Residual Stream - 16k
    5. 20-GEMMASCOPE-RES-16K
    6. 10251
    Prev
    Next
    INDEX
    Explanations

    phrases indicating potential for danger or risk, particularly in relation to gas and its effects

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

    clauses that describe a location or situation where something occurs, particularly involving subjects and verbs in sequences.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68
    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
    λευτα
    -0.35
    édrale
    -0.35
     مواليد
    -0.35
    Còn
    -0.35
    Reich
    -0.34
     Ajoutez
    -0.34
     this
    -0.34
    thebibliography
    -0.33
    entista
    -0.32
    isen
    -0.31
    POSITIVE LOGITS
     where
    0.73
     где
    0.72
    where
    0.70
     donde
    0.68
     ahol
    0.66
     ProtoMessage
    0.65
    NameInMap
    0.63
     όπου
    0.63
    InitStruct
    0.63
     Onde
    0.59
    Activations Density 0.219%

    No Known Activations