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

    Neuronpedia

    Jacobian LensNEW
    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. 29194
    Prev
    Next
    INDEX
    Explanations

    conditional statements and programming constructs related to logic flow

    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
    314
    +0.23
    1.1%
    745
    +0.13
    0.6%
    627
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    300
    +0.23
    0.04
    283
    +0.13
    0.01
    582
    +0.11
    0.03
    Negative Logits
     newcomers
    -0.61
     handwriting
    -0.60
     roadside
    -0.57
     Gren
    -0.57
     propri
    -0.56
     Ware
    -0.56
     brist
    -0.55
     that
    -0.54
     Solitaire
    -0.54
     barric
    -0.54
    POSITIVE LOGITS
    rame
    1.18
    rit
    0.98
    ihad
    0.96
    enth
    0.93
    ace
    0.91
    amily
    0.91
    acial
    0.90
    ram
    0.89
    ornia
    0.88
    ividual
    0.88
    Activations Density 0.370%

    No Known Activations