© 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
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Qwen2.5-7B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 75205
    Prev
    Next
    INDEX
    Explanations

    question marks

    np_max-act · gemini-2.0-flash

    This neuron detects interrogative sentences — it activates strongly at question endings (the question mark) and on question-like phrasing.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-qwen2.5-7b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     booty
    -0.07
    启示
    -0.07
    �
    -0.07
    宣传片
    -0.07
    �
    -0.07
    됄
    -0.07
    EPROM
    -0.07
    ト
    -0.06
     khỏe
    -0.06
    纹理
    -0.06
    POSITIVE LOGITS
     outlined
    0.08
    ⓥ
    0.07
    ഗ
    0.06
     rim
    0.06
    ulfilled
    0.06
     concise
    0.06
     nanop
    0.06
    jącej
    0.06
    .virtual
    0.06
     deadliest
    0.06
    Activations Density 0.077%

    No Known Activations