© 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
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 9057
    Prev
    Next
    INDEX
    Explanations

    topic headings and structured information within academic or technical texts

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

    beginning of document tokens (<bos>) and elements of webpage formatting like categories, external links, and headings.

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

    markers of document boundaries and structural metadata, especially Wikipedia-style section/category headers and the start-of-sequence token.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    the beginning of a document (start-of-sequence token).

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    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
    MouseOver
    -0.29
     вы
    -0.28
     fil
    -0.27
     interior
    -0.26
     фа
    -0.25
     Neuen
    -0.25
    場で
    -0.25
     use
    -0.24
     Description
    -0.24
     thừa
    -0.24
    POSITIVE LOGITS
    GEBURTSDATUM
    1.64
    Datuak
    1.59
    Portail
    1.52
     дописавши
    1.34
    expandindo
    1.32
     виправивши
    1.32
    ंदीखरीदारी
    1.30
    uxxxx
    1.27
    تقاوى
    1.26
    Personensuche
    1.26
    Activations Density 0.158%

    No Known Activations