© 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. 3-ATT-KK
    11. 14274
    Prev
    Next
    INDEX
    Explanations

    phrases related to investigations and integrity, especially in the context of governmental programs and information

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

    phrases that express the concept of integrity in various contexts

    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
    0:0.02
    1:0.02
    2:0.04
    3:0.24
    4:0.02
    5:0.03
    6:0.08
    7:0.18
    8:0.04
    9:0.04
    10:0.07
    11:0.16
    Negative Logits
    soDeliveryDate
    -1.31
    ycle
    -1.24
    姫
    -1.23
    KT
    -1.21
    jee
    -1.20
    urat
    -1.20
    undred
    -1.17
    heads
    -1.14
    icol
    -1.13
    emonium
    -1.12
    POSITIVE LOGITS
    rity
    1.35
     integrity
    1.35
     journalistic
    1.28
    tein
    1.25
    aminer
    1.21
     Remain
    1.18
     journalism
    1.16
    cellence
    1.15
     judgement
    1.14
     humanity
    1.14
    Activations Density 0.007%

    No Known Activations