© 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
    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. 13437
    Prev
    Next
    INDEX
    Explanations

    forum posts

    np_max-act · gemini-2.0-flash

    This neuron activates on the token “Not,” detecting when a user is negating or correcting something.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    detects the sentence-initial word "Not" (marking negation or contrast).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    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
     operates
    -0.07
    фи
    -0.06
    vens
    -0.06
    Allocation
    -0.06
    ुड
    -0.06
    raries
    -0.06
    出し
    -0.06
    ictions
    -0.05
     preceding
    -0.05
     rumored
    -0.05
    POSITIVE LOGITS
     editar
    0.08
     městě
    0.08
     زم
    0.07
     اشاره
    0.07
    �
    0.07
     Emb
    0.07
     sami
    0.07
     percentage
    0.07
     nem
    0.07
     chống
    0.06
    Activations Density 0.091%

    No Known Activations