© 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. 5-ATT-KK
    11. 37634
    Prev
    Next
    INDEX
    Explanations

    references to political conflicts and discussions surrounding them

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt
    Prompts (Dashboard)
    36,864 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    float32
    Hook Name
    blocks.5.attn.hook_z
    Hook Layer
    5
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.03
    1:0.03
    2:0.06
    3:0.13
    4:0.06
    5:0.08
    6:0.03
    7:0.03
    8:0.06
    9:0.19
    10:0.19
    11:0.06
    Negative Logits
     sidx
    -1.70
    assad
    -1.42
    ゴン
    -1.32
    ullah
    -1.29
    ajo
    -1.21
    yip
    -1.20
    ijah
    -1.17
     neighb
    -1.17
    kefeller
    -1.16
     quar
    -1.14
    POSITIVE LOGITS
    votes
    1.43
     jurors
    1.39
     hindsight
    1.30
     Feminist
    1.29
     commenters
    1.25
     libel
    1.22
     paraph
    1.21
    Truth
    1.21
     reddit
    1.21
     Gawker
    1.20
    Activations Density 1.239%

    No Known Activations