© 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. 8966
    Prev
    Next
    INDEX
    Explanations

    phrases related to offers and opportunities, especially involving rewards or positions

    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
    427
    +0.12
    0.6%
    369
    +0.12
    0.6%
    247
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.12
    0.04
    36
    +0.12
    0.04
    279
    +0.10
    0.03
    Negative Logits
    STER
    -0.78
    aha
    -0.67
    sen
    -0.67
    zl
    -0.66
     Already
    -0.66
    amazon
    -0.65
    umblr
    -0.65
    aneous
    -0.62
    _-
    -0.61
     Deaths
    -0.61
    POSITIVE LOGITS
     condolences
    1.10
     alternatives
    1.06
     congratulations
    1.00
     amnesty
    0.98
     refuge
    0.91
     unconditional
    0.89
     unres
    0.89
     alternative
    0.85
     forgiveness
    0.85
     mediation
    0.85
    Activations Density 0.299%

    No Known Activations