© 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. 8-RES_POST_32K-OAI
    11. 6916
    Prev
    Next
    INDEX
    Explanations

    bibliographic references and citations in academic or formal texts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    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
    373
    +0.19
    0.9%
    648
    +0.11
    0.5%
    561
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    373
    +0.19
    -0.02
    6
    +0.11
    0.04
    586
    +0.10
    0.03
    Negative Logits
     fortnight
    -0.63
    erous
    -0.62
    manship
    -0.60
     judiciary
    -0.60
    awed
    -0.60
     Derby
    -0.59
    orate
    -0.59
    elong
    -0.59
     2018
    -0.59
     favor
    -0.59
    POSITIVE LOGITS
     pp
    0.98
     pg
    0.85
     reprinted
    0.78
    itled
    0.78
    pp
    0.71
     cited
    0.69
     quoted
    0.69
     ISBN
    0.68
    Reviewer
    0.68
     pione
    0.68
    Activations Density 0.219%

    No Known Activations