© 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. 5834
    Prev
    Next
    INDEX
    Explanations

    instances of the phrase "to be" in various forms

    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
    +0.13
    0.6%
    373
    +0.13
    0.6%
    750
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.13
    0.04
    113
    +0.13
    0.03
    767
    +0.11
    0.03
    Negative Logits
     disapp
    -0.71
    audi
    -0.60
     indo
    -0.57
     encour
    -0.56
    forth
    -0.55
    ials
    -0.54
     obser
    -0.54
     denote
    -0.54
     proport
    -0.54
     mascul
    -0.53
    POSITIVE LOGITS
     grips
    1.14
    ubby
    0.81
    axy
    0.81
    asted
    0.74
    inces
    0.73
    usalem
    0.70
    eworld
    0.68
    ausible
    0.67
    ench
    0.66
    ivating
    0.66
    Activations Density 0.120%

    No Known Activations