© 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. 7741
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the absence or lack of something

    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
    669
    +0.13
    0.6%
    609
    +0.11
    0.5%
    234
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    669
    +0.13
    0.04
    455
    +0.11
    0.04
    609
    +0.10
    0.04
    Negative Logits
    raq
    -0.81
    geist
    -0.76
    =-=-=-=-=-=-=-=-
    -0.75
    çĶ°
    -0.74
    lish
    -0.73
    mun
    -0.72
    late
    -0.71
    ased
    -0.69
    qua
    -0.68
    aroo
    -0.68
    POSITIVE LOGITS
     mentioning
    1.10
     knowing
    1.09
     bothering
    1.04
     noticing
    1.04
     compromising
    1.03
     regard
    1.02
     sacrificing
    1.00
     realizing
    1.00
     interruption
    0.99
     committing
    0.97
    Activations Density 0.083%

    No Known Activations