© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 5-ATT-KK
    11. 17070
    Prev
    Next
    INDEX
    Explanations

    occurrences of specific forms of the word "cup" or related concepts of measuring and increase

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 5-att-kk
    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
    0:0.02
    1:0.01
    2:0.07
    3:0.07
    4:0.07
    5:0.04
    6:0.06
    7:0.37
    8:0.04
    9:0.03
    10:0.11
    11:0.06
    Negative Logits
    govtrack
    -1.77
     guiIcon
    -1.73
    ridor
    -1.73
    ickr
    -1.72
     INFO
    -1.70
    orescence
    -1.67
    inational
    -1.64
     istg
    -1.59
    isan
    -1.56
     [-
    -1.56
    POSITIVE LOGITS
     conditions
    1.90
     postponed
    1.77
     compromise
    1.77
     patched
    1.75
     condition
    1.75
     precedent
    1.66
     circumstance
    1.65
     unforeseen
    1.63
     circumstances
    1.63
     timing
    1.60
    Activations Density 0.000%

    No Known Activations