© 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. 126148
    Prev
    Next
    INDEX
    Explanations

    World

    np_max-act · gemini-2.0-flash

    The neuron fires on parenthetical stage‐direction fragments that introduce a character’s internal thought—e.g. “(to himself…)”

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

    parenthetical stage directions or asides indicating a character speaking to themself (e.g., “(to himself)”) and similar internal/action cues.

    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
     inflate
    -0.07
    -rise
    -0.07
     wash
    -0.06
    칙
    -0.06
     Orange
    -0.06
    Pressure
    -0.06
     isEqual
    -0.06
     vocals
    -0.06
    Link
    -0.06
     Others
    -0.06
    POSITIVE LOGITS
    	i
    0.07
    WhiteSpace
    0.07
     zveřej
    0.06
    �
    0.06
    entifier
    0.06
     Hurt
    0.06
     nhé
    0.06
     Όμιλος
    0.06
    бор
    0.06
     jazy
    0.06
    Activations Density 0.417%

    No Known Activations