© 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. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 9-ATT-KK
    11. 14210
    Prev
    Next
    INDEX
    Explanations

    phrases indicating potential negative consequences or risks

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 9-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    float32
    Hook Name
    blocks.9.attn.hook_z
    Hook Layer
    9
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.04
    1:0.02
    2:0.08
    3:0.20
    4:0.02
    5:0.08
    6:0.01
    7:0.08
    8:0.02
    9:0.01
    10:0.36
    11:0.02
    Negative Logits
    ��
    -2.55
    feat
    -1.95
     articulated
    -1.95
     Smooth
    -1.93
    basic
    -1.92
     articulate
    -1.89
     eloqu
    -1.87
     Loving
    -1.83
    raq
    -1.83
    phabet
    -1.83
    POSITIVE LOGITS
     unnecessarily
    3.48
     jeopard
    3.43
     worse
    3.42
     damaging
    3.42
     Worse
    3.39
     destabil
    3.23
     detriment
    3.22
     counterproductive
    3.19
     misunderstand
    3.18
     unintended
    3.14
    Activations Density 1.303%

    No Known Activations