© 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. 5277
    Prev
    Next
    INDEX
    Explanations

    instances of denial or restricted access being discussed in various contexts

    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
    204
    +0.15
    0.7%
    681
    +0.09
    0.4%
    467
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    204
    +0.15
    0.03
    567
    +0.09
    0.02
    234
    +0.09
    0.02
    Negative Logits
    inki
    -0.67
    ebin
    -0.64
    ipel
    -0.63
     behavi
    -0.63
     Recover
    -0.63
    Interstitial
    -0.62
    weeney
    -0.61
     Whitman
    -0.61
     dust
    -0.60
     Balanced
    -0.60
    POSITIVE LOGITS
     existence
    0.93
     permission
    0.91
     entry
    0.88
     privileges
    0.86
     access
    0.86
     request
    0.86
     requests
    0.86
     outright
    0.83
     exist
    0.83
     legitimacy
    0.81
    Activations Density 0.330%

    No Known Activations