© 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. Qwen3.5-4B
    3. 15-RES-MATRYOSHKA-65K
    4. 21857
    Prev
    Next
    INDEX
    Explanations

    , or file names like 'pic1.jpg'. * "JSON object with loops? The end JSON should look like item1: [ { "location": [ {"latitude": value, "longitude": value, "stopNum" * "user2 spain user3 italy ... File input 2: user1 trackname1 user2 trackname2 user3 trackname1 ... I need the number of users per country," * "I have three prize, each prize has one pic, like: prize1 pic1 prize2 pic2 prize3 pic3" * "formula field which will use a field from the first row. Example: (ColumnAFromCurrentRecord - ColumnBFromFirstRecord) / ColumnBFromFirstRecord" * "DayNo & CR_Date with data: Day1 01/01/2000 Day2 01/02/2"Let's synthesize these observations:* **MAX_ACTIVATING_TOKENS** highlights numerical values ('1') or potentially letter identifiers ('A').file name or attribute data* **TOKENS_AFTER_MAX_ACTIVATING

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    decoderesearch/qwen-3.5-saes/qwen-3.5-4b
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    oth
    -0.06
    ungs
    -0.06
     or
    -0.05
    oru
    -0.05
    auce
    -0.05
    forth
    -0.05
    åĬ¨æīĭèĥ½åĬĽ
    -0.05
    Ñħодов
    -0.05
    auge
    -0.05
     (.
    -0.05
    POSITIVE LOGITS
    دÙĬد
    0.07
    æĸĩåĵ²
    0.06
    åĮħçļĦ
    0.06
    esen
    0.06
    ernel
    0.06
     sod
    0.06
    åĪĨè§£
    0.05
     ronde
    0.05
    çĻ
    0.05
    çļĦä¸Ģ项æĺ¯
    0.05
    Activations Density 0.008%

    No Known Activations