© 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
    3. 31-GEMMASCOPE-2-RES-262K
    4. 122030
    Prev
    Next
    INDEX
    Explanations

    following details

    np_acts-logits-general · gemini-2.5-flash-lite

    election**Explanation Process:**1. **Analyze `<MAX_ACTIVATING_TOKENS>`:** `maximizing`, `one`, `discovering`, `sub`, `nomination`. These seem like keywords hinting at specific contexts.2. **Analyze `<TOKENS_AFTER_MAX_ACTIVATING_TOKEN>`:** `size` (after maximizing), `Aster` (after one), `sink` (after one), `sub` (after and)3. **Analyze `<TOP_ACTIVATING_TEXTS>`:** * "...**maximizing** size..." * "...Old Man Willow smiled. "Sometimes,"" (not very clear pattern) * "...* **Shape:** Asteroids aren't perfect spheres." (mentions "Aster", related to one of the tokens) * "...LOAD_BUFFER; end" (code context) * "...**discovering** what \*you\* and your partner(s) enjoy." * "...`[ebp-24h].message`, IDA doesn't know that `message` is a specific field within a structure. election

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-pt/resid_post/layer_31_width_262k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     سعودی
    0.58
    రోజు
    0.48
    ನಾಡ
    0.47
    番組
    0.46
    ಂತೆ
    0.45
    予以
    0.45
    젯
    0.45
    まるで
    0.45
    MEDIA
    0.44
    老師
    0.43
    POSITIVE LOGITS
    ou
    0.52
    iko
    0.51
     went
    0.48
    ome
    0.48
    meh
    0.46
    ม
    0.46
    EL
    0.46
    ulp
    0.45
    .},
    0.45
    ре
    0.45
    Activations Density 0.001%

    No Known Activations