© 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. 6785
    Prev
    Next
    INDEX
    Explanations

    terms associated with flawed or problematic systems and their implications

    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
    165
    +0.10
    0.4%
    101
    +0.10
    0.4%
    124
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.10
    0.03
    503
    +0.10
    0.02
    4
    +0.09
    0.03
    Negative Logits
     mum
    -0.69
     Lans
    -0.63
    ussion
    -0.61
    opsy
    -0.60
    VD
    -0.58
     Prel
    -0.58
    PDATE
    -0.57
    rush
    -0.57
     feedback
    -0.56
     Perspect
    -0.56
    POSITIVE LOGITS
    ably
    2.26
    able
    1.99
    ables
    1.64
    ability
    1.61
    ABLE
    1.47
    abilities
    1.34
    ibly
    1.29
    abl
    1.11
    ivable
    1.05
    ivably
    1.04
    Activations Density 0.186%

    No Known Activations