© 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. 8924
    Prev
    Next
    INDEX
    Explanations

    expressions of condolences and emotional responses to loss

    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
    487
    +0.11
    0.5%
    220
    +0.09
    0.4%
    486
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    566
    +0.11
    0.16
    354
    +0.09
    0.15
    718
    +0.09
    0.14
    Negative Logits
     [|
    -0.66
    Lear
    -0.61
    quin
    -0.61
    electric
    -0.61
    arton
    -0.60
    nesota
    -0.59
    ilogy
    -0.59
    acles
    -0.59
    obb
    -0.58
    hack
    -0.58
    POSITIVE LOGITS
     apologize
    1.21
     apologise
    1.19
     hereby
    1.10
     respectfully
    1.07
    hement
    1.01
     apologies
    1.00
     unres
    1.00
     honoured
    0.98
     congratulate
    0.96
     condol
    0.95
    Activations Density 1.255%

    No Known Activations