© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Attention SAE Research Paper
    5. GPT2-Small
    7. Attention Out
    9. 5-ATT-KK
    11. 42338
    Prev
    Next
    INDEX
    Explanations

    verb forms, particularly in the past tense

    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
    0:0.01
    1:0.01
    2:0.07
    3:0.06
    4:0.13
    5:0.02
    6:0.12
    7:0.37
    8:0.03
    9:0.03
    10:0.04
    11:0.05
    Negative Logits
    ussen
    -1.74
    onomous
    -1.74
    heses
    -1.68
    vironment
    -1.68
    unci
    -1.67
    apter
    -1.66
    soDeliveryDate
    -1.65
    rection
    -1.65
    adiq
    -1.58
    apest
    -1.58
    POSITIVE LOGITS
     Tears
    1.87
     excess
    1.82
     sweat
    1.81
     negligence
    1.73
     curiosity
    1.68
     havoc
    1.59
     tears
    1.56
     intimidation
    1.55
    gery
    1.54
     nostalgia
    1.52
    Activations Density 0.000%

    No Known Activations