© 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. 8-RES_POST_32K-OAI
    11. 7625
    Prev
    Next
    INDEX
    Explanations

    mentions of human rights violations and asylum issues

    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
    488
    +0.12
    0.5%
    50
    +0.10
    0.5%
    542
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    86
    +0.12
    0.11
    385
    +0.10
    0.16
    680
    +0.10
    0.02
    Negative Logits
    itialized
    -0.92
     Asgard
    -0.88
    odiac
    -0.85
     congr
    -0.84
     Hawking
    -0.84
     electrons
    -0.84
     Gohan
    -0.84
    Fans
    -0.78
     Genius
    -0.78
     Saiyan
    -0.77
    POSITIVE LOGITS
     prosecutions
    1.41
     Amnesty
    1.19
     abuses
    1.18
     impunity
    1.15
     audits
    1.15
     punitive
    1.14
     crackdown
    1.13
     NGO
    1.12
     prisons
    1.11
     enforcement
    1.11
    Activations Density 11.746%

    No Known Activations