© 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
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 41452
    Prev
    Next
    INDEX
    Explanations

    horror

    np_max-act · gemini-2.0-flash

    horror or creepy narrative elements.

    oai_token-act-pair · gpt-4oTriggered by @tcai

    The neuron fires on third‐person narrative segments in horror prose—spotting personal pronouns (he, him, his) and the surrounding verbs that describe bodily sensations or panic.

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

    elements related to supernatural horror, particularly themes involving ghosts, curses, and haunting experiences.

    oai_token-act-pair · gpt-4o-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
     sampled
    -0.07
    (:,
    -0.07
     คำ
    -0.06
    .ra
    -0.06
    .uniform
    -0.06
     бл
    -0.06
    FINAL
    -0.06
    但
    -0.06
    dim
    -0.06
    zzo
    -0.06
    POSITIVE LOGITS
    istency
    0.07
    流
    0.06
    _need
    0.06
     RAM
    0.06
     NSCoder
    0.06
     Savannah
    0.06
    createElement
    0.06
    .setError
    0.06
    °
    0.06
     Insurance
    0.06
    Activations Density 0.056%

    No Known Activations