© 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. 10-RES_POST_32K-OAI
    11. 31685
    Prev
    Next
    INDEX
    Explanations

    superlative adjectives indicating extremes or notable qualities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    756
    +0.11
    0.5%
    262
    +0.10
    0.5%
    162
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    122
    +0.11
    0.04
    194
    +0.10
    0.03
    262
    +0.10
    0.03
    Negative Logits
    yang
    -0.78
    ending
    -0.77
    iliar
    -0.74
    fixed
    -0.72
    alus
    -0.71
    skirts
    -0.71
    anism
    -0.70
    olor
    -0.68
    kas
    -0.68
     Cosponsors
    -0.68
    POSITIVE LOGITS
     possible
    1.46
     conceivable
    1.37
     imaginable
    1.32
     practicable
    1.13
     bidder
    0.91
     ever
    0.90
     feasible
    0.88
     thing
    0.87
     Possible
    0.84
    iary
    0.80
    Activations Density 0.281%

    No Known Activations