© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 6211
    Prev
    Next
    INDEX
    Explanations

    instances of the word "barely" and variations indicating minimal effort or presence

    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
    170
    +0.10
    0.5%
    129
    +0.10
    0.4%
    616
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    221
    +0.10
    0.02
    718
    +0.10
    0.02
    762
    +0.09
    0.02
    Negative Logits
    qua
    -0.67
     Parenthood
    -0.64
    È
    -0.64
     Preferred
    -0.63
     Helpful
    -0.61
    atu
    -0.61
    ãĤº
    -0.59
    isive
    -0.59
     Destination
    -0.59
    Reviewer
    -0.58
    POSITIVE LOGITS
     scraping
    0.96
     scrape
    0.87
     surviving
    0.86
     survived
    0.84
     tolerated
    0.82
     glimps
    0.82
     scratched
    0.81
     clinging
    0.79
     scratching
    0.79
     disguised
    0.79
    Activations Density 0.082%

    No Known Activations