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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B-IT
    3. 12-GEMMASCOPE-2-RES-16K
    4. 6067
    Prev
    Next
    INDEX
    Explanations

    **Detecting Numbers and Dates**I'm analyzing the provided lists to understand the specific patterns or concepts the neuron is most sensitive to. My goal is to explain this behavior in a concise phrase, focusing on potential meanings derived from digit sequences, words, or positional relationships.**Analysis Steps:**1. **MAX_ACTIVATING_TOKENS:** Contains sequences of digits: `3, 0, 3, 3, 9, 0, 9, 9, 9`. The digits `0`, `3`, and `9` are prominent.2. **TOKENS_AFTER_MAX_ACTIVATING_TOKEN:** Contains `2, ,, ,, 7, 0, 3, 9, 9`.**Detecting Numerical Sequences and Structure**I am currently examining the provided lists to identify underlying patterns. My aim is to synthesize these observations into a concise phrase that describes what the neuron detects or predicts based on numerical sequences, token positions, and contextual clues from activating texts.**Analysis based on provided lists:*** **MAX\_ACTIVATING\_TOKENS:** Contains primarily digits: `3, 0, 3, 3, 9, 0, 9, 9, 9`. This strongly suggests a numerical focus, with `0`, `3`, and `9` being frequent.* **TOKENS\_AFTER\_

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-it/resid_post/layer_12_width_16k_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
    ك
    1.95
     exons
    1.88
    тах
    1.76
    ного
    1.71
     hostilities
    1.67
     immunoglobulin
    1.59
     illustrator
    1.57
    trimethyl
    1.55
    lights
    1.52
     envo
    1.52
    POSITIVE LOGITS
    서
    1.85
    তিনি
    1.80
    ximately
    1.80
     Saltar
    1.77
    ić
    1.75
    $\$
    1.75
    관
    1.74
    Depuis
    1.73
    ó
    1.73
    ία
    1.70
    Activations Density 0.109%

    No Known Activations