© 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
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 73809
    Prev
    Next
    INDEX
    Explanations

    Illegality

    np_max-act · gemini-2.0-flash

    The neuron activates on adjectival or adverbial terms labeling wrongful or illegal conduct (e.g. “improper,” “unlawful,” “wrongfully,” “illegally”).

    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
    cete
    -0.06
     TARGET
    -0.06
     Alive
    -0.06
    baseline
    -0.06
    .Ref
    -0.06
     fence
    -0.06
     대부분
    -0.06
     能
    -0.06
     Analytics
    -0.06
     Loans
    -0.05
    POSITIVE LOGITS
    ATAR
    0.07
    lius
    0.07
    apatkan
    0.07
    .netflix
    0.07
    ừa
    0.07
    şam
    0.07
    cych
    0.06
    鄉
    0.06
     homelessness
    0.06
    conut
    0.06
    Activations Density 0.008%

    No Known Activations