© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 29722
    Prev
    Next
    INDEX
    Explanations

    references to interior characteristics and contexts

    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
    288
    +0.11
    0.5%
    637
    +0.09
    0.4%
    408
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.11
    0.03
    401
    +0.09
    0.02
    537
    +0.08
    0.01
    Negative Logits
    bilt
    -0.87
    tle
    -0.83
    nir
    -0.83
    astically
    -0.78
    tailed
    -0.77
    isSpecialOrderable
    -0.75
    lda
    -0.75
    thood
    -0.74
    urai
    -0.73
    atche
    -0.71
    POSITIVE LOGITS
     decoration
    0.93
     ministry
    0.92
    most
    0.91
     decor
    0.90
    ity
    0.88
    ITY
    0.81
     designer
    0.81
     Designer
    0.80
    ized
    0.79
    ITIES
    0.77
    Activations Density 0.101%

    No Known Activations