© 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. 10-RES_POST_32K-OAI
    11. 31512
    Prev
    Next
    INDEX
    Explanations

    occurrences of forms of the verb "to be" and related verbs indicating existence or action

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    89
    +0.11
    0.5%
    528
    +0.10
    0.5%
    749
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    7
    +0.11
    0.15
    102
    +0.10
    0.14
    55
    +0.10
    0.14
    Negative Logits
     Slug
    -0.68
     messenger
    -0.67
     Fighters
    -0.65
     Awakens
    -0.65
    isen
    -0.63
     nic
    -0.62
     Milk
    -0.61
     abusers
    -0.61
     syndrome
    -0.60
    bnb
    -0.59
    POSITIVE LOGITS
     undertaken
    1.58
     conducted
    1.48
     performed
    1.42
     underway
    1.34
     halted
    1.27
     done
    1.26
     carried
    1.25
     initiated
    1.24
     commenced
    1.20
     accomplished
    1.16
    Activations Density 1.023%

    No Known Activations