© 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
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Qwen2.5-7B-IT
    7. Resid Post - 131k
    9. 15-RESID-POST-AA
    11. 17461
    Prev
    Next
    INDEX
    Explanations

    writing and communication

    np_max-act · gemini-2.0-flash

    This neuron detects instruction-style prompt text — imperatives and metadata that ask the model to write or perform a specific task (e.g., "Write...", format/type labels, and nearby punctuation).

    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_15/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ett
    -0.07
     Alan
    -0.07
     københavn
    -0.06
    /*------------------------------------------------
    -0.06
     unto
    -0.06
     denne
    -0.06
    定向
    -0.06
    departureday
    -0.06
     Subway
    -0.06
    /example
    -0.06
    POSITIVE LOGITS
     변
    0.08
    时期的
    0.07
     &$
    0.07
     único
    0.07
    熬
    0.07
     الاجتماع
    0.07
    /gtest
    0.07
    קיימת
    0.06
     esl
    0.06
     sol
    0.06
    Activations Density 0.104%

    No Known Activations