© 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. 10167
    Prev
    Next
    INDEX
    Explanations

    complex relationships and dependencies in information

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

    phrases beginning with "via" followed by "the," particularly in descriptions of attack methods or techniques.

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

    constructions that introduce qualifying or explanatory subclauses—especially method/means phrases, relative clauses, and parenthetical asides.

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

    the word "which" when it introduces a clause that provides additional information or serves as a relative pronoun connecting clauses.

    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
    @[+][
    -0.53
    UnusedPrivate
    -0.51
    Ecotoxicity
    -0.49
    expandindo
    -0.48
    hoeddwyd
    -0.48
     Italijanski
    -0.47
     Foire
    -0.40
     ElevatedButton
    -0.39
     urg
    -0.39
    InstrumentedTest
    -0.39
    POSITIVE LOGITS
    +#+#
    0.59
    angliski
    0.46
    principalTable
    0.46
     nubes
    0.40
     llorando
    0.40
    ecore
    0.40
     Guille
    0.39
     anaknya
    0.38
    rungsseite
    0.38
     dorados
    0.38
    Activations Density 0.737%

    No Known Activations