© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 4269
    Prev
    Next
    INDEX
    Explanations

    the word "more" and its variants, indicating a focus on additional information or quantity

    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
    314
    +0.12
    0.5%
    647
    +0.11
    0.5%
    454
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    726
    +0.12
    0.07
    389
    +0.11
    0.07
    190
    +0.10
    0.07
    Negative Logits
    DragonMagazine
    -0.85
    theless
    -0.68
     Rebell
    -0.64
     Dunk
    -0.63
    going
    -0.63
    luaj
    -0.63
     Watching
    -0.60
    ween
    -0.59
    spring
    -0.58
    taking
    -0.58
    POSITIVE LOGITS
     details
    0.88
     detailed
    0.87
     detail
    0.85
     precise
    0.80
     formally
    0.78
    requently
    0.78
     contiguous
    0.78
     exact
    0.75
     simplified
    0.75
     efficiently
    0.72
    Activations Density 0.337%

    No Known Activations