© 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. 7293
    Prev
    Next
    INDEX
    Explanations

    phrases related to attention and focus on specific topics

    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
    674
    +0.14
    0.7%
    337
    +0.12
    0.5%
    91
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.14
    0.14
    669
    +0.12
    0.08
    589
    +0.10
    0.07
    Negative Logits
    emale
    -0.90
    phia
    -0.81
    idem
    -0.75
    strike
    -0.73
    amen
    -0.72
    azo
    -0.71
    assadors
    -0.70
    rection
    -0.67
    rights
    -0.67
    ieves
    -0.66
    POSITIVE LOGITS
     gimm
    1.16
     distractions
    1.08
     frivol
    1.08
     frivolous
    1.07
     clich
    1.06
     flashy
    1.03
     fancy
    1.01
     fanc
    1.00
     overly
    0.99
     boring
    0.97
    Activations Density 2.153%

    No Known Activations