© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 13099
    Prev
    Next
    INDEX
    Explanations

    negative or error-related phrases

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/4-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    348
    +0.14
    0.8%
    78
    +0.14
    0.8%
    47
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    47
    +0.14
    0.03
    461
    +0.14
    0.03
    36
    +0.14
    0.03
    Negative Logits
     conservation
    -1.49
     reality
    -1.48
     concerned
    -1.44
     issue
    -1.42
     affairs
    -1.42
     carriers
    -1.39
     concern
    -1.31
     manifestation
    -1.30
     correctly
    -1.29
     kindly
    -1.29
    POSITIVE LOGITS
    ³
    3.35
    ı
    2.94
    Ĥ
    2.87
    ½
    2.87
    ĭ
    2.81
    ·
    2.80
    İ
    2.79
    ¼
    2.77
    Ĺ
    2.73
    Ĩ
    2.72
    Activations Density 0.191%

    No Known Activations