© 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. 5749
    Prev
    Next
    INDEX
    Explanations

    infinitive forms of verbs starting with "to."

    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
    718
    +0.09
    0.4%
    194
    +0.09
    0.4%
    251
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    718
    +0.09
    0.07
    49
    +0.09
    0.07
    608
    +0.09
    0.06
    Negative Logits
    ERG
    -0.75
     teasp
    -0.71
    terday
    -0.71
    Hor
    -0.69
    plates
    -0.65
    minus
    -0.64
     Needs
    -0.61
    emonium
    -0.61
    imental
    -0.60
     Flags
    -0.59
    POSITIVE LOGITS
     stomach
    1.15
     comprehend
    1.09
     discern
    1.07
     decipher
    1.05
     convince
    1.03
     reconcil
    0.99
     grasp
    0.97
     justify
    0.96
     convincing
    0.95
     reconcile
    0.92
    Activations Density 0.207%

    No Known Activations