© 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 EvalsExportsAPI 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. 6476
    Prev
    Next
    INDEX
    Explanations

    themes related to authority, truthfulness, and societal critique

    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
    762
    +0.15
    0.7%
    18
    +0.11
    0.5%
    241
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    762
    +0.15
    0.18
    225
    +0.11
    0.15
    288
    +0.11
    0.17
    Negative Logits
    trak
    -0.78
    assic
    -0.77
    ãĤ©
    -0.67
    Rail
    -0.66
    lav
    -0.65
     confir
    -0.64
    UNCH
    -0.63
    joining
    -0.63
    ouple
    -0.63
    emp
    -0.63
    POSITIVE LOGITS
     lest
    1.11
     unnecessarily
    0.93
     whilst
    0.92
     blindly
    0.86
     rather
    0.85
     instead
    0.85
     freely
    0.84
     oneself
    0.82
     amidst
    0.82
     uncond
    0.82
    Activations Density 2.356%

    No Known Activations