Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 55706
    Prev
    Next
    INDEX
    Explanations

    is

    np_max-act · gemini-2.0-flash

    The neuron is highly sensitive to formal legal/administrative terminology—words naming official roles, obligations, or procedures.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    deictic and discourse‑marker terms that point to or frame information within the text (e.g., demonstratives, existential “there,” and contrast/citation cues).

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
     пер
    -0.08
     اف
    -0.07
     Schneider
    -0.06
    .Dispatcher
    -0.06
    虽然
    -0.06
    ddd
    -0.06
     patrol
    -0.06
    451
    -0.06
    construction
    -0.06
     cont
    -0.06
    POSITIVE LOGITS
    _MODIFIED
    0.06
    ิทธ
    0.06
    を見
    0.06
    icot
    0.06
    ovid
    0.06
     голод
    0.06
     maneu
    0.06
     aquatic
    0.06
     Bankası
    0.06
    学
    0.06
    Activations Density 0.195%

    No Known Activations