© 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. 4912
    Prev
    Next
    INDEX
    Explanations

    negative terms or concepts, particularly in the context of technology or competition

    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
    85
    +0.11
    0.5%
    312
    +0.11
    0.5%
    619
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    537
    +0.11
    0.04
    425
    +0.11
    0.03
    152
    +0.09
    0.03
    Negative Logits
    ians
    -0.61
     Otto
    -0.61
    arium
    -0.61
    aceae
    -0.59
     Boxing
    -0.59
     Zar
    -0.59
    ano
    -0.59
     lest
    -0.58
     Bland
    -0.58
     Guth
    -0.57
    POSITIVE LOGITS
    purpose
    1.11
    handedly
    1.00
    dimensional
    0.94
    species
    0.94
    sided
    0.93
    channel
    0.93
    edged
    0.92
    colored
    0.90
    disciplinary
    0.90
    task
    0.90
    Activations Density 0.043%

    No Known Activations