© 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-IT
    7. Residual Stream - 16k
    9. 20-GEMMASCOPE-RES-16K
    11. 8850
    Prev
    Next
    INDEX
    Explanations

    references to programs and their characteristics

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

    "program"

    np_acts-logits-general · gemini-2.0-flash

    programs and their use

    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
    wet
    -0.47
    white
    -0.45
    isDead
    -0.43
     Bélgica
    -0.42
     detecta
    -0.41
     Kills
    -0.41
     Melhor
    -0.41
    wild
    -0.41
     gosta
    -0.40
     Úl
    -0.40
    POSITIVE LOGITS
     Program
    1.25
     program
    1.20
    Program
    1.16
     PROGRAM
    1.13
     Programs
    1.12
    PROGRAM
    1.10
    Programs
    1.09
     Programme
    1.08
     programs
    1.04
    sprogramm
    1.03
    Activations Density 0.057%

    No Known Activations