© 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. 7717
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "on" and related prepositions or phrases indicating actions or events

    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
    163
    +0.10
    0.5%
    346
    +0.09
    0.4%
    686
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    686
    +0.10
    0.03
    471
    +0.09
    0.03
    553
    +0.08
    0.03
    Negative Logits
    drawn
    -0.95
    uid
    -0.89
     wrinkles
    -0.85
    cius
    -0.79
    vec
    -0.78
    zl
    -0.77
    erred
    -0.75
    hander
    -0.73
    rence
    -0.73
    lied
    -0.72
    POSITIVE LOGITS
     civilians
    0.95
     Kabul
    0.92
     Pearl
    0.91
     embassies
    0.87
     innocent
    0.86
     civilian
    0.85
     Samar
    0.83
     Tours
    0.82
     unarmed
    0.81
     Baghdad
    0.80
    Activations Density 0.354%

    No Known Activations