© 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. 8-RES_POST_32K-OAI
    11. 5175
    Prev
    Next
    INDEX
    Explanations

    instances of differentiation and distinctions among concepts or categories

    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
    658
    +0.10
    0.5%
    109
    +0.10
    0.5%
    221
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    575
    +0.10
    0.07
    724
    +0.10
    0.07
    495
    +0.10
    0.04
    Negative Logits
     earnest
    -0.68
     excess
    -0.65
     understatement
    -0.62
    ccoli
    -0.58
    upid
    -0.56
     underrated
    -0.56
     omission
    -0.55
    fin
    -0.55
    unc
    -0.54
     absence
    -0.54
    POSITIVE LOGITS
     altogether
    1.02
     than
    0.95
     depending
    0.95
    than
    0.92
    iates
    0.88
    iating
    0.87
    iations
    0.85
     different
    0.83
    Different
    0.81
    styles
    0.78
    Activations Density 0.749%

    No Known Activations