© 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. 5041
    Prev
    Next
    INDEX
    Explanations

    phrases indicating a strong presence or emphasis on the concept of "a" in various 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
    105
    +0.10
    0.5%
    548
    +0.10
    0.4%
    596
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.10
    0.22
    362
    +0.10
    0.13
    608
    +0.09
    0.11
    Negative Logits
    osponsors
    -0.77
    terms
    -0.76
    姫
    -0.75
    votes
    -0.72
    deals
    -0.72
    tests
    -0.71
     Reasons
    -0.70
     Facts
    -0.69
    Reason
    -0.69
    ividual
    -0.69
    POSITIVE LOGITS
     eerie
    1.19
     dazzling
    1.10
     lush
    1.10
     exqu
    1.08
     sleek
    1.07
     surreal
    1.06
     blazing
    1.04
     fiery
    1.02
     strikingly
    1.00
     ferocious
    0.99
    Activations Density 1.124%

    No Known Activations