© 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. 18797
    Prev
    Next
    INDEX
    Explanations

    refusing or refusal

    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
    ç»ĿæľĽ
    -0.09
    è¾ĵäºĨ
    -0.09
    çĥ§
    -0.09
    åĬłæ²¹
    -0.09
    unlock
    -0.09
    fail
    -0.08
    newsletter
    -0.08
     unlocked
    -0.08
     Piet
    -0.08
    æĵĶ
    -0.08
    POSITIVE LOGITS
    æĭĴç»Ŀ
    0.15
    æĭĴçµķ
    0.13
    æĭĴ
    0.12
    åĪ¶æŃ¢
    0.11
    客æ°Ķ
    0.11
    éĺ²çº¿
    0.10
    åĿļåĨ³
    0.10
     firm
    0.10
     refusing
    0.10
     refusal
    0.10
    Activations Density 0.056%

    No Known Activations