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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 5434
    Prev
    Next
    INDEX
    Explanations

    frequencies and intensities of the word "most" and its variations

    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
    373
    +0.13
    0.6%
    696
    +0.09
    0.4%
    494
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    190
    +0.13
    0.06
    45
    +0.09
    0.03
    102
    +0.09
    0.05
    Negative Logits
     messenger
    -0.68
     Wonderful
    -0.61
     trump
    -0.59
    BuyableInstoreAndOnline
    -0.59
     feminine
    -0.58
     dod
    -0.57
     antique
    -0.57
     harassing
    -0.56
     Trance
    -0.56
     intangible
    -0.56
    POSITIVE LOGITS
    ially
    0.96
    probably
    0.94
    usually
    0.92
     certainly
    0.90
    often
    0.89
    inally
    0.89
    then
    0.89
    inas
    0.85
    typically
    0.83
     likely
    0.81
    Activations Density 0.265%

    No Known Activations