© 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. 129660
    Prev
    Next
    INDEX
    Explanations

    Research/Technical Abstracts

    np_max-act · gemini-2.0-flash

    This neuron detects technical mentions of “weight” or “weighted” (and related qualifiers like unweighted, pseudo-polynomial) in graph/algorithm descriptions.

    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
     brain
    -0.07
     deco
    -0.07
     FIXME
    -0.07
     tele
    -0.07
     minecraft
    -0.07
     swimming
    -0.07
    our
    -0.07
     notifier
    -0.06
     implements
    -0.06
     texting
    -0.06
    POSITIVE LOGITS
    ersh
    0.07
    -heavy
    0.07
    although
    0.06
     fruity
    0.06
    BASE
    0.06
    _province
    0.06
    rapid
    0.06
    еліг
    0.06
    한국
    0.06
     хотел
    0.06
    Activations Density 0.261%

    No Known Activations