© 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
    7. Residual Stream - 16k
    9. 31-GEMMASCOPE-RES-16K
    11. 7901
    Prev
    Next
    INDEX
    Explanations

    specific scientific terms and jargon, particularly related to biochemistry and medical conditions

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

    medical and scientific terminology, particularly related to diseases, medical conditions, and biological research.

    oai_token-act-pair · claude-3-5-haiku-20241022Triggered by @neilrathi

    scientific and technical terminology, particularly abbreviated terms or specialized vocabulary.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    technical and scientific terminology, particularly multi-syllabic compound words and specialized domain vocabulary.

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

    words that start with a capital letter (proper nouns, sentence beginnings, or acronyms).

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

    capitalized word prefixes in long, technical or formal terms (e.g., section headings, scientific/legal terminology, and acronyms).

    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
     ModelExpression
    -0.60
    ########.
    -0.58
     rhestr
    -0.56
    Chham
    -0.54
    يكب
    -0.53
    httphttps
    -0.53
     незавершена
    -0.51
    Personensuche
    -0.51
     بيها
    -0.50
    Transkript
    -0.49
    POSITIVE LOGITS
     EconPapers
    0.48
    hogy
    0.46
     ModelAndView
    0.41
     '${
    0.41
    Lalu
    0.39
    Bew
    0.39
    lihood
    0.38
    FieldBuilder
    0.38
    Beaver
    0.37
     BoxFit
    0.37
    Activations Density 0.648%

    No Known Activations