© 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. 7665
    Prev
    Next
    INDEX
    Explanations

    intensifiers or adjectives that express strong emotions or emphasis

    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
    125
    +0.13
    0.6%
    185
    +0.10
    0.5%
    155
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.13
    0.07
    55
    +0.10
    0.07
    540
    +0.10
    0.06
    Negative Logits
    rongh
    -0.66
    rings
    -0.65
    senal
    -0.65
    abouts
    -0.64
    umbnail
    -0.63
     Odyssey
    -0.62
    umbn
    -0.62
    ertodd
    -0.61
    appings
    -0.61
    ukong
    -0.60
    POSITIVE LOGITS
     raining
    1.30
     impossible
    1.12
     easier
    1.02
     possible
    1.00
     advisable
    0.96
     easy
    0.94
     ironic
    0.94
     conceivable
    0.94
     unclear
    0.93
    ceivable
    0.91
    Activations Density 0.247%

    No Known Activations