© 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. 30795
    Prev
    Next
    INDEX
    Explanations

    phrases related to observation or attention

    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
    1
    +0.11
    0.5%
    334
    +0.10
    0.4%
    129
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    750
    +0.11
    0.03
    526
    +0.10
    -0.01
    382
    +0.09
    0.02
    Negative Logits
    ĪĴ
    -0.69
    Constructed
    -0.62
    chairs
    -0.61
    stood
    -0.61
    cribed
    -0.59
    ramid
    -0.59
     moder
    -0.59
     Mandal
    -0.59
     Atkins
    -0.58
    rium
    -0.57
    POSITIVE LOGITS
     glimpse
    1.14
     whiff
    1.06
     attention
    0.96
     glimps
    0.92
     breath
    0.90
     drift
    0.88
    eye
    0.85
     fleeting
    0.83
    ById
    0.80
     scent
    0.77
    Activations Density 0.623%

    No Known Activations