© 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
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 4822
    Prev
    Next
    INDEX
    Explanations

    instances of the word "use."

    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
    314
    +0.15
    0.7%
    540
    +0.12
    0.6%
    380
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    318
    +0.15
    0.03
    329
    +0.12
    0.03
    146
    +0.11
    0.03
    Negative Logits
     Buk
    -0.68
     Sag
    -0.68
     Brill
    -0.68
     Kirin
    -0.68
     Nas
    -0.66
     Scouting
    -0.63
     Patron
    -0.63
     passage
    -0.62
     sailing
    -0.62
     moderation
    -0.62
    POSITIVE LOGITS
    ptions
    0.96
    able
    0.96
    ername
    0.95
     thereof
    0.88
    ables
    0.86
    ability
    0.84
    fulness
    0.82
    lessness
    0.81
    full
    0.80
    mone
    0.79
    Activations Density 0.104%

    No Known Activations