© 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. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 8842
    Prev
    Next
    INDEX
    Explanations

    phrases related to business, management, and penalties

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.14
    0.6%
    282
    +0.10
    0.4%
    231
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.16
    575
    +0.10
    0.13
    762
    +0.10
    0.13
    Negative Logits
    stood
    -0.74
     confir
    -0.67
     hemor
    -0.66
    ãĤ»
    -0.62
    ounced
    -0.61
    UGE
    -0.59
    iren
    -0.57
     Merlin
    -0.57
    RET
    -0.56
     pestic
    -0.56
    POSITIVE LOGITS
     entails
    1.25
     requires
    1.20
     helps
    1.12
     involves
    1.07
     implies
    1.01
     reduces
    0.99
     improves
    0.99
     means
    0.99
     oneself
    0.99
     gives
    0.97
    Activations Density 3.430%

    No Known Activations