© 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. 2-GEMMASCOPE-RES-16K
    11. 13820
    Prev
    Next
    INDEX
    Explanations

    the concept of "slipping" or "twisting" in various contexts, particularly related to failure or deception, as well as references to "poison."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 2-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     المعيارى
    -0.66
    səhifə
    -0.60
    URBANA
    -0.54
    DialogContent
    -0.51
    Datuak
    -0.51
    ьажоргаш
    -0.51
    HasAnnotation
    -0.50
    idenav
    -0.50
    Portail
    -0.50
     EconPapers
    -0.48
    POSITIVE LOGITS
     poison
    0.74
     poisoning
    0.73
     Poison
    0.68
    Poison
    0.67
     slip
    0.62
     exploit
    0.62
    slip
    0.60
     twist
    0.58
    Slip
    0.57
    poison
    0.56
    Activations Density 0.291%

    No Known Activations