© 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. 5430
    Prev
    Next
    INDEX
    Explanations

    negative qualities or conditions associated with people, environments, or societal 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
    362
    +0.20
    0.9%
    408
    +0.09
    0.4%
    680
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.20
    0.14
    408
    +0.09
    0.14
    288
    +0.09
    0.13
    Negative Logits
    utics
    -0.86
    hots
    -0.81
     besides
    -0.81
    ;;
    -0.78
    ovember
    -0.76
    helm
    -0.76
    each
    -0.73
     resembling
    -0.72
     Finally
    -0.70
     indeed
    -0.69
    POSITIVE LOGITS
     nature
    1.23
     portion
    1.21
    est
    1.19
    liest
    1.12
     confines
    1.08
     assumption
    1.07
     aspect
    1.06
     aftermath
    1.05
     notion
    1.01
     workings
    1.01
    Activations Density 1.353%

    No Known Activations