© 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. 7683
    Prev
    Next
    INDEX
    Explanations

    contradictory statements and expressions that highlight dissenting opinions

    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
    205
    +0.11
    0.5%
    689
    +0.10
    0.4%
    218
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    8
    +0.11
    0.09
    598
    +0.10
    0.05
    14
    +0.10
    0.05
    Negative Logits
     cancellation
    -0.68
     Pwr
    -0.66
     Priority
    -0.63
     disqualified
    -0.60
     waning
    -0.60
     attrition
    -0.60
     leftover
    -0.59
    icultural
    -0.59
     residual
    -0.59
    cellaneous
    -0.59
    POSITIVE LOGITS
     actually
    1.03
    actually
    0.96
    Actually
    0.95
    agine
    0.93
     Actually
    0.89
     reality
    0.89
     truth
    0.88
    reality
    0.80
     actual
    0.79
    APS
    0.67
    Activations Density 2.029%

    No Known Activations