© 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 EvalsExportsinterp-engineNEWAPI 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. 7414
    Prev
    Next
    INDEX
    Explanations

    programming constructs related to templates and allocation functions

    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
    373
    +0.45
    2.3%
    265
    +0.11
    0.6%
    640
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.45
    0.08
    701
    +0.11
    0.05
    160
    +0.09
    0.08
    Negative Logits
     describ
    -0.71
     sorely
    -0.65
     IMAGES
    -0.60
     smack
    -0.59
     agre
    -0.59
     ASAP
    -0.58
    ALSE
    -0.58
     thereby
    -0.58
     murd
    -0.58
     deval
    -0.57
    POSITIVE LOGITS
    prototype
    0.91
    exit
    0.77
    entry
    0.75
    illon
    0.72
    \":
    0.68
    fix
    0.67
    adr
    0.66
    outer
    0.66
    irin
    0.66
    operator
    0.65
    Activations Density 0.678%

    No Known Activations