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

    Neuronpedia

    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. 12140
    Prev
    Next
    INDEX
    Explanations

    references to specific issues or subjects within a broader context

    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
    156
    +0.18
    1.0%
    348
    +0.17
    1.0%
    199
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    14
    +0.18
    0.10
    359
    +0.17
    0.10
    237
    +0.12
    0.07
    Negative Logits
    ħ
    -2.21
    º
    -1.93
    ĭ
    -1.91
    ·¸
    -1.84
    Ń
    -1.84
    °
    -1.77
    ĥ½
    -1.70
    ¬
    -1.68
    ı
    -1.66
    Ļª
    -1.64
    POSITIVE LOGITS
     kind
    2.31
     particular
    2.15
     elusive
    2.15
     same
    2.10
     phenomenon
    2.08
     caliber
    2.02
     endeavor
    2.01
     situation
    2.00
     sort
    1.99
     latter
    1.97
    Activations Density 0.578%

    No Known Activations