© 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. 28902
    Prev
    Next
    INDEX
    Explanations

    instances of evaluation or commentary on experiences and perceptions

    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
    763
    +0.14
    0.7%
    218
    +0.10
    0.5%
    478
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    720
    +0.14
    0.18
    546
    +0.10
    0.12
    43
    +0.10
    0.11
    Negative Logits
     NOT
    -0.82
    rones
    -0.76
    ļéĨĴ
    -0.71
    LU
    -0.68
    æ©Ł
    -0.68
    rather
    -0.67
    nt
    -0.67
    instead
    -0.65
    Not
    -0.65
    not
    -0.64
    POSITIVE LOGITS
     depending
    1.30
    depending
    1.26
     anytime
    0.83
     someday
    0.82
     Either
    0.79
    Either
    0.78
     or
    0.72
     if
    0.71
     indulge
    0.69
     succumb
    0.68
    Activations Density 1.095%

    No Known Activations