© 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. 8243
    Prev
    Next
    INDEX
    Explanations

    references to screens or screen content

    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
    360
    +0.11
    0.5%
    167
    +0.10
    0.4%
    504
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    678
    +0.11
    0.02
    76
    +0.10
    0.02
    519
    +0.09
    0.02
    Negative Logits
    doms
    -0.72
    ortium
    -0.69
    aternal
    -0.68
     plur
    -0.68
     Carnegie
    -0.61
    shire
    -0.60
    vernment
    -0.59
    akeru
    -0.59
    ghan
    -0.58
     caution
    -0.58
    POSITIVE LOGITS
    plays
    1.33
    writers
    1.09
    writer
    1.01
    printed
    0.99
    writing
    0.98
    grab
    0.97
    caps
    0.93
    casts
    0.88
    cheat
    0.88
    aver
    0.86
    Activations Density 0.056%

    No Known Activations