© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4440
    Prev
    Next
    INDEX
    Explanations

    instances of punctuation marks, particularly apostrophes used in contractions or possessives

    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
    477
    +0.12
    0.5%
    373
    +0.10
    0.5%
    459
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    701
    +0.12
    0.01
    764
    +0.10
    0.01
    508
    +0.09
    0.01
    Negative Logits
     Annotations
    -0.84
    lihood
    -0.76
     ner
    -0.69
    âĢİ
    -0.63
    netflix
    -0.60
    itiz
    -0.60
    ovember
    -0.60
     resear
    -0.58
     Nib
    -0.57
    nia
    -0.57
    POSITIVE LOGITS
    thumbnails
    1.16
    taboola
    1.01
    Interstitial
    0.96
    inventoryQuantity
    0.80
    POST
    0.77
    });
    0.76
    mix
    0.71
    submit
    0.71
    ItemThumbnailImage
    0.64
    umbn
    0.64
    Activations Density 0.025%

    No Known Activations