© 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. 20-GEMMASCOPE-RES-16K
    11. 11288
    Prev
    Next
    INDEX
    Explanations

    phrases related to specific locations and features in a narrative context

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

    high-frequency function words (articles, pronouns, auxiliaries) and possessive apostrophes, often at sentence starts or immediately before nouns/dates.

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

    Looking at the activations, this neuron activates strongly on the word "the" when it appears before nouns in formal or technical contexts, particularly in descriptive passages about objects, locations, or processes. The neuron is finding **the definite article "the" in formal,

    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
     either
    -0.46
    Personendaten
    -0.45
     but
    -0.41
    ValueStyle
    -0.41
     or
    -0.40
     chaînes
    -0.40
     businesses
    -0.39
     temporarily
    -0.39
    ineries
    -0.39
     strategies
    -0.38
    POSITIVE LOGITS
     kohdetta
    0.61
    CppMethod
    0.54
    featureID
    0.54
    يكب
    0.52
    Xna
    0.49
    PROTOBUF
    0.48
    SBATCH
    0.47
     /*#__
    0.47
    PYX
    0.46
    UnitTesting
    0.46
    Activations Density 0.241%

    No Known Activations