© 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. 44216
    Prev
    Next
    INDEX
    Explanations

    narrative text

    np_max-act · gemini-2.0-flash

    This neuron detects tokens in the role‐playing or prompt‐injection instructions—i.e. the meta text that tells the model what character to play and how to behave.

    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
    ,dim
    -0.07
    .fp
    -0.07
    justify
    -0.06
    .bind
    -0.06
     drawn
    -0.06
    _bind
    -0.06
    -powered
    -0.06
    _PLAYER
    -0.06
    zin
    -0.06
     Pes
    -0.06
    POSITIVE LOGITS
     XCTAssertEqual
    0.07
    tele
    0.07
     krás
    0.07
    msgid
    0.07
    іж
    0.06
    assertInstanceOf
    0.06
    _exempt
    0.06
     Guy
    0.06
    皆
    0.06
     Corbyn
    0.06
    Activations Density 0.024%

    No Known Activations