© 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. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 15-RESID-POST-AA
    11. 56279
    Prev
    Next
    INDEX
    Explanations

    Formal documents and articles

    np_max-act · gemini-2.0-flash

    It detects self-referential assistant utterances—tokens in which the model speaks about itself (first-person "I" / "as an AI" style statements).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-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
    рин
    -0.07
    -hand
    -0.07
    iconductor
    -0.07
    ,new
    -0.07
     Mid
    -0.06
     Ya
    -0.06
    atsapp
    -0.06
    GRAM
    -0.06
    BinContent
    -0.06
     तरफ
    -0.06
    POSITIVE LOGITS
    FD
    0.07
     выдел
    0.06
    elsea
    0.06
     Collision
    0.06
    /oct
    0.06
     الحياة
    0.06
    (draw
    0.06
     �
    0.06
     그녀
    0.05
     jr
    0.05
    Activations Density 0.194%

    No Known Activations