© 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. 10-RES_POST_32K-OAI
    11. 26532
    Prev
    Next
    INDEX
    Explanations

    the repetition of the word "more" and variations that express an increase or enhancement

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    681
    +0.10
    0.5%
    253
    +0.09
    0.4%
    411
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    726
    +0.10
    0.04
    181
    +0.09
    0.04
    132
    +0.09
    0.03
    Negative Logits
    interstitial
    -0.76
    ilts
    -0.68
     CoC
    -0.67
     DRAG
    -0.64
    unknown
    -0.64
     understatement
    -0.61
    verson
    -0.60
     Difference
    -0.59
    ubs
    -0.57
    going
    -0.57
    POSITIVE LOGITS
     vigorously
    1.33
     effectively
    1.28
     forcefully
    1.28
     efficiently
    1.27
     harshly
    1.25
     decisively
    1.24
     aggressively
    1.24
     thoroughly
    1.21
     confidently
    1.19
     freely
    1.18
    Activations Density 0.114%

    No Known Activations