© 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. 9-GEMMASCOPE-RES-131K
    11. 78482
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate communication and interaction

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

    explaining or saying something

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

    This neuron detects direct instructions or advice phrased to the reader (imperatives and second‑person directive language).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k
    Configuration
    google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.9.hook_resid_post
    Hook Layer
    9
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    parametrize
    -0.36
     ibunya
    -0.36
     nødven
    -0.36
     inférieure
    -0.35
    Apr
    -0.34
     klubben
    -0.34
     wzi
    -0.32
    時点で
    -0.32
     hjelp
    -0.32
     vč
    -0.31
    POSITIVE LOGITS
     autorytatywna
    0.54
    KommentareTeilen
    0.54
     setuptools
    0.53
    afficheront
    0.51
    queryInterface
    0.48
    SerializedSize
    0.47
    ArrowToggle
    0.46
    SharedCtor
    0.46
    دانشنامهٔ
    0.45
    tvguidetime
    0.44
    Activations Density 0.356%

    No Known Activations