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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B-IT
    3. 12-GEMMASCOPE-2-RES-16K
    4. 3487
    Prev
    Next
    INDEX
    Explanations

    This neuron seems to be associated with past tense verbs, particularly those used in compound tenses (like "have been", "have taken"). The pattern it detects is likely the formation of past actions or states.Considering the provided lists:- **MAX_ACTIVATING_TOKENS**: `have`, `ve`, `ve`, `d`, `d`, `ve`, `ve`, `ve`, `ve`, `have`. This strongly suggests auxiliary verbs like "have", contracted "have" (like "I've"), or the auxiliary "had".- **TOKENS_AFTER_MAX_ACTIVATING_TOKEN**: `drifted`, `been`, `been`, `built`, `imagined`, `taken`, `instructed`, `had`, `provided`, `been`. These are all past participles or past tense verbs. They commonly follow the tokens in `MAX_ACTIVATING_TOKENS` (e.g., "have drifted", "have been", "have built", "have taken", "have instructed", "have provided").past actions or completed states

    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
    が多く
    1.30
     zakres
    1.26
    юць
    1.19
     dentées
    1.12
    asambhavam
    1.09
     Такой
    1.05
    CollectedField
    1.05
    рија
    1.04
     podido
    1.02
    ளமான
    1.01
    POSITIVE LOGITS
    á
    1.20
    (
    1.11
    m
    1.09
    Di
    1.03
    م
    1.00
    ม
    1.00
    to
    0.99
    nn
    0.96
    ない
    0.95
    ương
    0.95
    Activations Density 0.026%

    No Known Activations