© 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. 4841
    Prev
    Next
    INDEX
    Explanations

    conjunctions and phrases that indicate relationships among various elements

    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
    544
    +0.12
    0.6%
    71
    +0.12
    0.6%
    410
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    393
    +0.12
    0.13
    678
    +0.12
    0.08
    738
    +0.10
    0.06
    Negative Logits
    aeda
    -0.87
    Technology
    -0.85
     Authors
    -0.80
    Science
    -0.79
    ylum
    -0.79
    Services
    -0.79
     Occupations
    -0.79
    wives
    -0.78
     Anthropology
    -0.77
    States
    -0.77
    POSITIVE LOGITS
     tumble
    1.18
     swipe
    1.15
     jerk
    1.12
     pinch
    1.07
     slur
    1.05
     shuffle
    1.05
     flo
    1.05
     flick
    1.04
     shove
    1.03
     scram
    1.03
    Activations Density 1.968%

    No Known Activations