© 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. 11-RESID-POST-AA
    11. 30133
    Prev
    Next
    INDEX
    Explanations

    personal anecdotes

    np_max-act · gemini-2.0-flash

    The neuron is detecting demonstrative/deictic words or phrases (e.g. “de där dagarna” in Swedish or “that … way” in English).

    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
     كيل
    -0.06
    uu
    -0.06
    圣
    -0.06
    )%
    -0.06
    加工
    -0.06
     pev
    -0.06
     inviting
    -0.06
    レス
    -0.06
    afc
    -0.06
     extensively
    -0.06
    POSITIVE LOGITS
    QS
    0.07
     ГО
    0.06
     dejtingsaj
    0.06
     něm
    0.06
    Ess
    0.06
    �
    0.06
     GameState
    0.06
    жения
    0.06
     неж
    0.06
     Eş
    0.06
    Activations Density 0.113%

    No Known Activations