© 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. 21495
    Prev
    Next
    INDEX
    Explanations

    phrases indicating positive or negative developments or trends over time

    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
    2:0.23
    3:0.11
    4:0.16
    5:0.02
    6:0.05
    7:0.11
    8:0.04
    9:0.03
    10:0.09
    11:0.09
    Negative Logits
     retains
    -1.48
    emark
    -1.42
     subscrib
    -1.41
    20439
    -1.36
     denies
    -1.31
     qualifies
    -1.30
    assis
    -1.30
     identifies
    -1.30
    oho
    -1.29
    catentry
    -1.28
    POSITIVE LOGITS
    spir
    1.46
    Plot
    1.22
     Gaw
    1.21
     worse
    1.20
     Weir
    1.20
     deterioration
    1.17
    oward
    1.17
     vengeance
    1.16
     downhill
    1.15
     Grind
    1.15
    Activations Density 0.040%

    No Known Activations