© 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. 11-RESID-POST-AA
    11. 58656
    Prev
    Next
    INDEX
    Explanations

    2

    np_max-act · gemini-2.0-flash

    This neuron detects structural or meta-text elements — instructions, labels, formatted Q/A/JSON headings, numbers and other layout/formatting tokens used to organize content.

    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
    新规
    -0.07
     Maxim
    -0.07
    Seattle
    -0.07
    典型
    -0.07
    _sim
    -0.07
    创新
    -0.07
     simulations
    -0.07
     Scouts
    -0.07
     hepatitis
    -0.07
    visitor
    -0.07
    POSITIVE LOGITS
     Without
    0.08
    Without
    0.07
    —all
    0.07
    —you
    0.07
    AGE
    0.07
    emento
    0.06
    ▂
    0.06
     Until
    0.06
    𝙑
    0.06
    'T
    0.06
    Activations Density 1.080%

    No Known Activations