© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 5016
    Prev
    Next
    INDEX
    Explanations

    statements about the state of existence and societal issues related to people's experiences and perceptions

    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
    44
    +0.10
    0.4%
    420
    +0.09
    0.4%
    431
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    102
    +0.10
    0.17
    566
    +0.09
    0.12
    7
    +0.09
    0.12
    Negative Logits
    ounters
    -0.72
    ounter
    -0.72
    wered
    -0.71
    obos
    -0.69
    ilk
    -0.69
     sidx
    -0.68
    erred
    -0.68
    oji
    -0.67
    Joined
    -0.66
    bury
    -0.65
    POSITIVE LOGITS
     notoriously
    1.10
     finite
    0.99
     inherently
    0.97
     imperfect
    0.90
     understandably
    0.88
     complex
    0.85
     complicated
    0.83
     enormously
    0.82
     intrinsically
    0.81
     undeniably
    0.81
    Activations Density 2.060%

    No Known Activations