© 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. 3-ATT-KK
    11. 12807
    Prev
    Next
    INDEX
    Explanations

    adjectives and phrases surrounding praise or criticism

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    quoted speech or expressions of sentiment

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-att-kk
    Configuration
    ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L3_Hcat_z_lr1.20e-03_l19.00e-01_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.3.attn.hook_z
    Hook Layer
    3
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Head Attr Weights
    0:0.06
    1:0.07
    2:0.15
    3:0.03
    4:0.05
    5:0.11
    6:0.08
    7:0.08
    8:0.08
    9:0.06
    10:0.10
    11:0.08
    Negative Logits
    VERTISEMENT
    -1.48
     heights
    -1.29
    ership
    -1.28
     Crew
    -1.27
     <-
    -1.25
     sails
    -1.25
     craw
    -1.25
     guiActiveUn
    -1.22
    cies
    -1.18
     Desmond
    -1.16
    POSITIVE LOGITS
    connected
    1.33
    Magn
    1.33
    orthy
    1.32
    done
    1.31
    govtrack
    1.30
    rather
    1.28
    appropriately
    1.28
    lain
    1.22
    weak
    1.22
    ensitive
    1.20
    Activations Density 0.013%

    No Known Activations