© 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. 40517
    Prev
    Next
    INDEX
    Explanations

    conditional phrases indicating potential actions or possibilities

    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
    0:0.06
    1:0.04
    2:0.05
    3:0.09
    4:0.07
    5:0.08
    6:0.08
    7:0.09
    8:0.10
    9:0.08
    10:0.06
    11:0.16
    Negative Logits
    [/
    -1.65
     insur
    -1.40
    /+
    -1.35
     unlocks
    -1.34
    flation
    -1.33
     perk
    -1.32
    ][/
    -1.32
     Wars
    -1.32
    )]
    -1.31
    $.
    -1.31
    POSITIVE LOGITS
    ��
    1.66
     discharged
    1.48
    ディ
    1.44
    アル
    1.44
    ッド
    1.43
    Report
    1.36
    ミ
    1.36
    intention
    1.36
    Published
    1.33
     recite
    1.33
    Activations Density 0.005%

    No Known Activations