© 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. 6346
    Prev
    Next
    INDEX
    Explanations

    references to a character named "she" in various contexts

    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
    167
    +0.13
    0.6%
    606
    +0.12
    0.6%
    612
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    139
    +0.13
    0.09
    454
    +0.12
    0.08
    224
    +0.12
    0.07
    Negative Logits
    kefeller
    -0.75
     Leban
    -0.72
     Seym
    -0.70
     srfAttach
    -0.69
    perties
    -0.66
    ornia
    -0.65
    hovah
    -0.62
     Jindal
    -0.61
    INGTON
    -0.61
     Disclosure
    -0.59
    POSITIVE LOGITS
    pher
    1.50
    athed
    1.24
    pard
    1.19
    lled
    1.16
    ppard
    1.08
    athing
    1.06
    ffield
    1.05
    pherd
    1.04
    ding
    1.04
    ldon
    1.03
    Activations Density 0.285%

    No Known Activations