© 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
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 91106
    Prev
    Next
    INDEX
    Explanations

    Stating an opinion

    np_max-act · gemini-2.0-flash

    This neuron detects hedge or disclaimer phrases—especially first-person “I’m not…” or “not suggesting…” style qualifiers.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-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
     Sir
    -0.07
     recycled
    -0.07
    .previous
    -0.06
    _requests
    -0.06
     Vish
    -0.06
    ebilir
    -0.06
    mh
    -0.06
     female
    -0.06
     Wells
    -0.06
     percentage
    -0.06
    POSITIVE LOGITS
     �
    0.07
     bàn
    0.06
    .Route
    0.06
    _EXTRA
    0.06
     ノ
    0.06
    adjust
    0.06
     식
    0.06
    Snake
    0.06
    (Size
    0.06
     trainable
    0.06
    Activations Density 0.029%

    No Known Activations