© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 32657
    Prev
    Next
    INDEX
    Explanations

    instances of the verb "tell" and its variations in the context of narrating stories or conveying information

    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
    94
    +0.10
    0.5%
    314
    +0.10
    0.4%
    373
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    750
    +0.10
    0.04
    230
    +0.10
    0.04
    327
    +0.10
    0.04
    Negative Logits
     Frieza
    -0.67
     dies
    -0.59
    nant
    -0.58
    nikov
    -0.57
    cill
    -0.56
     fixme
    -0.56
    aila
    -0.56
     disappear
    -0.55
     memos
    -0.55
    locks
    -0.55
    POSITIVE LOGITS
     us
    0.94
     readers
    0.81
    ¿½
    0.76
     reader
    0.73
    tale
    0.72
     viewer
    0.72
     viewers
    0.71
    Ĥª
    0.69
    ingly
    0.66
    reader
    0.64
    Activations Density 0.731%

    No Known Activations