© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4930
    Prev
    Next
    INDEX
    Explanations

    the occurrence and variations of the word "Gr" or "Br" in different contexts

    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
    23
    +0.13
    0.6%
    709
    +0.10
    0.4%
    356
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.13
    0.04
    314
    +0.10
    0.00
    60
    +0.10
    0.03
    Negative Logits
    ĸļ
    -0.81
    Thumbnail
    -0.73
     obsolete
    -0.66
    ciation
    -0.66
     Vide
    -0.63
     Scenes
    -0.63
    âĸ¬âĸ¬
    -0.63
     Pastebin
    -0.62
     lights
    -0.60
    Extra
    -0.60
    POSITIVE LOGITS
    ace
    1.00
    ady
    0.86
    icial
    0.84
    icit
    0.83
    izz
    0.82
    acy
    0.82
    illing
    0.81
    illy
    0.81
    anie
    0.81
    ynes
    0.79
    Activations Density 0.133%

    No Known Activations