© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-27B-IT
    3. 15-GEMMASCOPE-2-TRANSCODER-262K
    4. 14627
    Prev
    Next
    INDEX
    Explanations

    This is challenging because the "TOP_POSITIVE_LOGITS" list seems disconnected from the other lists. The other lists strongly suggest themes of network layers, applications, enterprises, and DDoS mitigation.Let's re-evaluate based on the strongest signals:- **MAX_ACTIVATING_TOKENS**: layer, enterprises, level, Applications, funds, Enterprise, application, User, Layer- **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: traffic, DDoS, Protocols, Defined, ProtocolsThese heavily point towards network security, specifically regarding applications and layers.application layer protocols

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_15_width_262k_l0_small_affine
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     lên
    0.47
     anunció
    0.47
     anunciou
    0.45
     tuyên
    0.42
     recordó
    0.42
     rebate
    0.41
     વિવિધ
    0.41
     V
    0.41
     العربية
    0.40
     anunciado
    0.40
    POSITIVE LOGITS
    waters
    0.45
     waters
    0.43
    管道
    0.42
     basins
    0.42
    populations
    0.42
    AppCompatTheme
    0.41
     jim
    0.41
    నికి
    0.41
     islands
    0.41
    quarters
    0.40
    Activations Density 0.066%

    No Known Activations