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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    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. 91453
    Prev
    Next
    INDEX
    Explanations

    Instructions/recipes

    np_max-act · gemini-2.0-flash

    The neuron activates on cooking‐recipe instruction steps, especially imperative verbs and measurement phrases (e.g., “Add,” “mix,” quantities and units).

    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
     permitted
    -0.06
    …and
    -0.06
     Carl
    -0.06
     Mol
    -0.06
    }while
    -0.05
     країни
    -0.05
    model
    -0.05
    'all
    -0.05
     accompanied
    -0.05
     clarification
    -0.05
    POSITIVE LOGITS
     richness
    0.07
    ν
    0.07
     equipment
    0.07
    RUN
    0.07
    (_("
    0.07
    �n
    0.06
    κ
    0.06
    ��
    0.06
    olución
    0.06
    starting
    0.06
    Activations Density 0.014%

    No Known Activations