© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-27B-IT
    3. 10-GEMMASCOPE-2-TRANSCODER-262K
    4. 207252
    Prev
    Next
    INDEX
    Explanations

    **TOP_ACTIVATING_TEXTS**: * `width="match_parent" android:layout_height="match_parent" android:src="@drawable/your_image" android:scaleType="centerCrop"/> <TextView android:layout` - Contains `@drawable/`.The neuron seems to be looking for things related to `drawable` and `ab`.The `MAX_ACTIVATING_TOKENS` are dominated by `drawable` and `ab`.The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` often show characters or short tokens that might follow these words in code or text. For `drawable`, we see `/`, `item`, and `,`. For `ab`, we see `ox`.The `TOP_ACTIVATING_TEXTS` also show examples of `@drawable/` and other contexts where `drawable` might appear (like `Drawable` enum in Rust, Drawabox). There's also content related to "ab" in the context of "ab-testing" or similar, and "ab" as a prefix in words like "about". However, the strongest signal is `drawable`.Let's re-examine the structure:- `drawable` is followed by `/` or `item`. This strongly suggests XML or similar markup where drawables are referenced (e.g., `@drawable/icon`).- `ab` is followed by `ox`. This is less clear. Could it be part of words like "box"?- `Drawable` (capitalized) is followed by `` ` ``.- `Rotating` is followed by `an`.- `Metrics` is followed by `().`.- `of` is followed by `JSON`.The `TOP_POSITIVE_LOGITS` don't provide a clear thematic link.The most consistent and prominent pattern is related to `drawable` and its usage in contexts like XML (`@drawable/`) or as an object/type name (`Drawable`). The `ab` pattern is also present but less clear what it signifies in isolation.Given the dominance of `drawable` and the appearance of `/` and `item` after it, a strong possibility is things related to referencing drawables in UI definitions. The `ab` might be related to "about" or other words.The request asks for a *concise* explanation (3-20 words).Let's try to combine the ideas:- `drawable /`- `drawable item`- `ab ox`The `drawable` pattern is stronger.`drawable /` or `ab ox`

    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_10_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
    器
    0.75
    زيد
    0.68
    ли
    0.66
    ↵
    0.63
    ون
    0.63
    ════
    0.63
    macht
    0.60
     cobrar
    0.59
    craft
    0.59
     arroz
    0.58
    POSITIVE LOGITS
    s
    0.89
    i
    0.85
    r
    0.83
    ی
    0.83
    IL
    0.72
    b
    0.72
    IX
    0.72
    AT
    0.71
    ORE
    0.71
    OR
    0.69
    Activations Density 0.000%

    No Known Activations