© 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. 31-GEMMASCOPE-2-RES-262K
    4. 110445
    Prev
    Next
    INDEX
    Explanations

    model parameters and CPUThe `MAX_ACTIVATING_TOKENS` list contains "Parameters" and "parameters" twice, and "CPU" once.The `TOP_ACTIVATING_TEXTS` extensively discuss model "parameters" (e.g., "number of parameters", "billions of parameters", "model size", "knobs the model adjusts", "parameters are, very roughly, the 'knobs' the model adjusts during training to learn patterns") and also mention "CPU" ("A decent CPU is needed", "The CPU handles the processing").The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` shows words like "handles" following "CPU".The `TOP_POSITIVE_LOGITS` are a mix, but "zahl", "zahlreiche", "countless" implies quantities, which parameters and CPU relate to.Considering the most prominent and recurring themes, "parameters" and "CPU" are strong candidates. The phrase "parameters and CPU" is concise and covers these key elements. parameters and CPU

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_31_width_262k_l0_medium
    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
     consensus
    0.49
     frontier
    0.48
     courteous
    0.47
     trainee
    0.46
     якому
    0.45
     uniformed
    0.45
     jurisdictional
    0.44
     lured
    0.44
     foreman
    0.44
    læ
    0.44
    POSITIVE LOGITS
     zahl
    0.51
     zahlreiche
    0.45
    spannung
    0.44
     makarna
    0.44
    頩
    0.43
    无数
    0.42
    설
    0.42
    Musik
    0.42
    钢
    0.42
    íš
    0.41
    Activations Density 0.006%

    No Known Activations