© 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. 126390
    Prev
    Next
    INDEX
    Explanations

    improvements, changes, opportunities

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on high-impact, attention-grabbing content words—especially key thematic or promotional nouns and verbs (e.g. “boost,” “call,” “flavor,” “adaptation”) that carry the main message.

    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
     handlers
    -0.08
     listener
    -0.06
    Persistent
    -0.06
    試
    -0.06
     durumunda
    -0.06
    .getRequestDispatcher
    -0.06
    (validate
    -0.06
    arLayout
    -0.06
    etect
    -0.06
     전
    -0.06
    POSITIVE LOGITS
    素
    0.07
    CHASE
    0.06
    imachinery
    0.06
    .Empty
    0.06
    един
    0.06
    ічного
    0.06
    }
↵
↵
↵
↵
    0.06
    tone
    0.06
     hep
    0.06
     philosopher
    0.06
    Activations Density 0.119%

    No Known Activations