© 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. 2-RES-SM
    11. 10059
    Prev
    Next
    INDEX
    Explanations

    references to tablets and their specifications

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    386
    +0.15
    0.9%
    87
    +0.13
    0.8%
    156
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    148
    +0.15
    0.00
    150
    +0.13
    0.02
    386
    +0.13
    0.00
    Negative Logits
    ĺ
    -2.01
    Ļª
    -1.89
    ©
    -1.85
    ¬
    -1.83
    ķ
    -1.78
    ĨĴ
    -1.77
    «
    -1.76
    ½
    -1.68
    ®
    -1.65
    ı
    -1.61
    POSITIVE LOGITS
    ynthetic
    1.73
    optic
    1.69
    chitz
    1.67
    accharide
    1.60
    nik
    1.53
    ential
    1.47
    ledge
    1.46
    parse
    1.46
    pper
    1.44
    sover
    1.42
    Activations Density 0.206%

    No Known Activations