© 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 EvalsExportsinterp-engineNEWAPI 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. 7327
    Prev
    Next
    INDEX
    Explanations

    technical terms and systematic descriptions related to processes or methods

    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
    +0.17
    0.8%
    671
    +0.11
    0.5%
    514
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    674
    +0.17
    0.05
    718
    +0.11
    0.08
    111
    +0.10
    0.03
    Negative Logits
     Leilan
    -0.91
    lost
    -0.74
    ¶ħ
    -0.72
     Haunted
    -0.69
    bda
    -0.69
    Freedom
    -0.68
    owship
    -0.68
    esity
    -0.67
    edience
    -0.67
    arenthood
    -0.65
    POSITIVE LOGITS
     Fig
    0.91
    Fig
    0.90
     labelled
    0.78
     indicate
    0.75
     note
    0.75
    again
    0.74
     say
    0.73
     specify
    0.72
     pictured
    0.71
     indicated
    0.71
    Activations Density 2.346%

    No Known Activations