© 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 - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 270
    Prev
    Next
    INDEX
    Explanations

    phrases and structures commonly used in formal or instructional writing, such as describing steps, providing explanations, or giving guidance.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @aml7990

    phrases related to academic rigor and methodology

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

    followed by prepositions or conjunctions

    np_acts-logits-general · gemini-2.0-flash

    concepts, conclusions, particular details

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91
    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
     betweenstory
    -0.58
    +#+
    -0.52
    pidou
    -0.48
    IUrlHelper
    -0.48
    passphrase
    -0.47
    genodigd
    -0.46
    ViewInit
    -0.46
    беріга
    -0.46
    contentLoaded
    -0.45
    bukkit
    -0.45
    POSITIVE LOGITS
    AndEndTag
    0.49
     narr
    0.44
     purpoſe
    0.38
     quantitative
    0.37
     Anſ
    0.37
     juſ
    0.35
    utilisons
    0.35
     abstrato
    0.34
     ſtate
    0.33
     inapropiados
    0.33
    Activations Density 0.358%

    No Known Activations