© 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 EvalsExportsAPI 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. 6631
    Prev
    Next
    INDEX
    Explanations

    instances of the word "better" and phrases suggesting improvement or urgency

    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
    +0.14
    0.6%
    560
    +0.11
    0.5%
    2
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    49
    +0.14
    0.02
    455
    +0.11
    0.02
    13
    +0.09
    0.01
    Negative Logits
    cible
    -0.60
     Niet
    -0.58
     adverse
    -0.57
    eur
    -0.57
    venue
    -0.56
     Manit
    -0.56
     passer
    -0.55
     Associated
    -0.54
    doors
    -0.54
     manifold
    -0.53
    POSITIVE LOGITS
     hurry
    1.06
     beware
    0.96
     pray
    0.89
     refrain
    0.83
     HAVE
    0.80
     prove
    0.80
     rethink
    0.79
     leave
    0.78
     be
    0.78
     prepare
    0.77
    Activations Density 0.071%

    No Known Activations