© 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. Qwen2.5-7B-IT
    7. Resid Post - 131k
    9. 15-RESID-POST-AA
    11. 26031
    Prev
    Next
    INDEX
    Explanations

    technical writing

    np_max-act · gemini-2.0-flash

    the neuron detects passive‑voice / past‑tense constructions (instances of "was" or similar + past participle) typical of formal technical or legal prose.

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-qwen2.5-7b-instruct/resid_post_layer_15/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.15.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     escorts
    -0.08
    ġ
    -0.08
    pk
    -0.07
    _chunks
    -0.07
    collector
    -0.07
    _sb
    -0.07
    +v
    -0.07
     Tiny
    -0.06
    liğini
    -0.06
    娇
    -0.06
    POSITIVE LOGITS
    	re
    0.08
     Banking
    0.07
    бир
    0.07
    alc
    0.07
     silver
    0.07
    �
    0.07
     iframe
    0.07
    ycastle
    0.07
    Wall
    0.07
     отдел
    0.06
    Activations Density 1.109%

    No Known Activations