© 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. 13403
    Prev
    Next
    INDEX
    Explanations

    avoidance or limitations

    np_max-act · gemini-2.0-flash

    The neuron fires on tokens in recommendation or advice phrases (e.g. “Yes…fine to…,” “Just be mindful…,” “Best to leave out…,” “Try adding…”)—in other words, it spots when the text is offering a suggestion or directive.

    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
    .Scheme
    -0.07
    -described
    -0.06
    _BU
    -0.06
    _spi
    -0.06
    Professor
    -0.06
    有人
    -0.06
    _Page
    -0.06
     ballot
    -0.06
    AccessException
    -0.05
     effort
    -0.05
    POSITIVE LOGITS
    835
    0.07
    601
    0.07
    451
    0.07
    637
    0.06
     beforeEach
    0.06
     Visa
    0.06
     кос
    0.06
    :</
    0.06
     expanding
    0.06
     Emerson
    0.06
    Activations Density 0.003%

    No Known Activations