© 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
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 5358
    Prev
    Next
    INDEX
    Explanations

    occurrences of the verb "make" and its 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
    190
    +0.13
    0.6%
    151
    +0.09
    0.4%
    433
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    402
    +0.13
    0.07
    12
    +0.09
    0.04
    366
    +0.09
    0.04
    Negative Logits
    ²¾
    -0.74
    Untitled
    -0.67
     <@
    -0.67
     Norn
    -0.67
    plain
    -0.66
    posts
    -0.66
    Ñı
    -0.65
     Mood
    -0.64
    âĸĪâĸĪâĸĪâĸĪ
    -0.64
     precept
    -0.63
    POSITIVE LOGITS
    elta
    0.72
     Ramos
    0.70
    zens
    0.70
    henko
    0.69
    emouth
    0.68
    enhagen
    0.68
     Shots
    0.67
     clutch
    0.66
     Univ
    0.64
    aceous
    0.62
    Activations Density 0.335%

    No Known Activations