© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 48100
    Prev
    Next
    INDEX
    Explanations

    toxic, unhealthy, or healthy words

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    adamkarvonen/qwen3-32b-saes/saes_Qwen_Qwen3-32B_batch_top_k/resid_post_layer_32
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ellen
    -0.10
    åĩºå¤Ħ
    -0.10
    æıĸ
    -0.10
    棹
    -0.09
    adir
    -0.09
    åıijå±ķä¸Ń
    -0.08
    æĹĮ
    -0.08
    åĿĢ
    -0.08
    大家åĪĨ享
    -0.08
    _registers
    -0.08
    POSITIVE LOGITS
    çĭŃ
    0.09
    åĭĸ
    0.09
    unded
    0.08
    çĥ«
    0.08
    è¿Ļ份
    0.08
    éĹ¯
    0.08
     Evelyn
    0.08
    utt
    0.08
    èĩªå¾ĭ
    0.08
    phe
    0.08
    Activations Density 0.001%

    No Known Activations