© 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. 4839
    Prev
    Next
    INDEX
    Explanations

    comparative phrases indicating something is better or worse than another

    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
    323
    +0.11
    0.5%
    350
    +0.11
    0.5%
    572
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.11
    0.04
    539
    +0.11
    0.03
    538
    +0.10
    0.03
    Negative Logits
    ãĤ¼ãĤ¦ãĤ¹
    -0.74
    acters
    -0.71
    ourse
    -0.69
    Activity
    -0.69
    isms
    -0.68
    Tip
    -0.67
    autical
    -0.67
    LIN
    -0.66
     Topics
    -0.66
    antry
    -0.66
    POSITIVE LOGITS
     crap
    0.85
     rotten
    0.85
     useless
    0.80
     mediocre
    0.74
     crappy
    0.71
     prem
    0.70
     lousy
    0.69
     nothing
    0.69
     nowhere
    0.69
     unbeat
    0.68
    Activations Density 0.324%

    No Known Activations