© 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. 30661
    Prev
    Next
    INDEX
    Explanations

    conditional phrases indicating choices or alternatives

    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
    400
    +0.12
    0.5%
    239
    +0.10
    0.5%
    280
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    393
    +0.12
    0.06
    449
    +0.10
    0.03
    239
    +0.10
    0.03
    Negative Logits
     Flavoring
    -0.75
     Travels
    -0.69
    CRIP
    -0.68
    neau
    -0.68
    ombat
    -0.64
    isol
    -0.61
    ipel
    -0.61
     Rocks
    -0.61
    WORK
    -0.59
    uple
    -0.57
    POSITIVE LOGITS
    rogen
    0.84
    rists
    0.83
    rea
    0.80
    udicrous
    0.78
    soType
    0.76
    RO
    0.75
     obligation
    0.74
    ro
    0.73
     reluct
    0.72
     expectation
    0.69
    Activations Density 0.507%

    No Known Activations