© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 5692
    Prev
    Next
    INDEX
    Explanations

    issues related to auditing and election processes

    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
    670
    +0.09
    0.4%
    694
    +0.09
    0.4%
    300
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.09
    0.24
    408
    +0.09
    0.20
    40
    +0.09
    0.11
    Negative Logits
     apocalypse
    -0.84
     tomorrow
    -0.83
    ↵Âł
    -0.81
     ;)
    -0.81
    morrow
    -0.80
     :)
    -0.79
    ciation
    -0.79
    iku
    -0.78
     ðŁĻĤ
    -0.75
     congr
    -0.74
    POSITIVE LOGITS
     improperly
    1.34
     unlawfully
    1.11
     subcontract
    1.10
     improper
    1.09
     inappropriately
    1.06
     billing
    1.04
     allegedly
    1.01
     supervisors
    1.01
     defendants
    1.00
     violations
    0.99
    Activations Density 8.625%

    No Known Activations