Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    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. 125649
    Prev
    Next
    INDEX
    Explanations

    prepositions

    np_max-act · gemini-2.0-flash

    This neuron responds to directional adverbial particles (e.g. up, down, out, off) in phrasal‐verb constructions.

    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
     为
    -0.07
    CE
    -0.07
    oyo
    -0.06
    ACE
    -0.06
     descriptor
    -0.06
    nce
    -0.06
    04
    -0.06
    itionally
    -0.06
     Territory
    -0.06
    成了
    -0.06
    POSITIVE LOGITS
     up
    0.08
    の中
    0.08
     out
    0.07
     down
    0.07
     inşa
    0.07
     Země
    0.07
     واب
    0.07
    livé
    0.06
     univerz
    0.06
     Up
    0.06
    Activations Density 0.079%

    No Known Activations