© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 10-RES_POST_32K-OAI
    6. 29273
    Prev
    Next
    INDEX
    Explanations

    phrases and actions related to investigation and movement

    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
    177
    +0.10
    0.5%
    464
    +0.10
    0.4%
    497
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    377
    +0.10
    0.17
    651
    +0.10
    0.14
    698
    +0.09
    0.14
    Negative Logits
    )}
    -0.79
    osponsors
    -0.62
    VERTISEMENT
    -0.61
    asca
    -0.59
    taboola
    -0.58
    Interstitial
    -0.57
     consequ
    -0.57
    includes
    -0.55
    ============
    -0.55
    %);
    -0.54
    POSITIVE LOGITS
     and
    1.12
     expecting
    0.83
     intending
    0.79
    and
    0.77
     AND
    0.75
     pretending
    0.75
     And
    0.69
    And
    0.68
     hoping
    0.67
     saying
    0.67
    Activations Density 6.419%

    No Known Activations