© 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. 47174
    Prev
    Next
    INDEX
    Explanations

    instances of the word "harness" and related variations, indicating a focus on empowerment or utilization of resources

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Head Attr Weights
    0:0.02
    1:0.00
    2:0.05
    3:0.05
    4:0.13
    5:0.02
    6:0.04
    7:0.43
    8:0.03
    9:0.04
    10:0.06
    11:0.06
    Negative Logits
    emetery
    -1.73
    dule
    -1.67
     condol
    -1.65
    irm
    -1.63
    dated
    -1.54
    merce
    -1.54
    rats
    -1.53
    mates
    -1.50
    Deploy
    -1.45
    emies
    -1.44
    POSITIVE LOGITS
     energy
    1.65
     supernatural
    1.61
     energies
    1.53
     power
    1.48
     accelerated
    1.44
    energy
    1.42
     momentum
    1.41
     powerful
    1.34
     psychic
    1.32
     initiative
    1.31
    Activations Density 0.002%

    No Known Activations