© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 9125
    Prev
    Next
    INDEX
    Explanations

    expressions related to the concept of "the thing" often followed by a commentary or observation about situations or problems

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    373
    +0.13
    0.6%
    265
    +0.10
    0.5%
    369
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    737
    +0.13
    0.06
    676
    +0.10
    0.05
    195
    +0.10
    0.03
    Negative Logits
    SHIP
    -0.70
    avement
    -0.65
    artney
    -0.63
    ittees
    -0.62
    arantine
    -0.61
    20439
    -0.58
     Annotations
    -0.56
    GBT
    -0.56
    Italy
    -0.55
     inflicting
    -0.55
    POSITIVE LOGITS
     caveat
    0.97
     truth
    0.85
     kicker
    0.85
     thing
    0.85
     downside
    0.78
     problem
    0.77
     takeaway
    0.75
     lesson
    0.72
     gist
    0.68
     nutshell
    0.68
    Activations Density 0.512%

    No Known Activations