© 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
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 5609
    Prev
    Next
    INDEX
    Explanations

    negative connotations or themes related to violence and harm

    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
    420
    +0.11
    0.5%
    11
    +0.10
    0.5%
    83
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    152
    +0.11
    0.08
    314
    +0.10
    0.02
    525
    +0.10
    0.05
    Negative Logits
     proportional
    -0.77
     adequate
    -0.75
     expend
    -0.75
     optimum
    -0.74
     discretionary
    -0.74
     reactive
    -0.74
     unsustainable
    -0.74
     stabilized
    -0.73
     inverse
    -0.72
     optimal
    -0.72
    POSITIVE LOGITS
    warts
    1.12
    lore
    1.11
    maker
    1.07
    dogs
    1.07
    worms
    1.06
    flies
    1.04
    girls
    1.02
    shaped
    1.02
    tale
    1.01
    fires
    1.00
    Activations Density 0.186%

    No Known Activations