© 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. 10143
    Prev
    Next
    INDEX
    Explanations

    topics related to technology and programming, especially concerning applications and systems

    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
    156
    +0.32
    2.0%
    369
    +0.19
    1.2%
    263
    +0.17
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.32
    0.03
    448
    +0.19
    0.03
    293
    +0.17
    0.07
    Negative Logits
    deal
    -1.66
    ately
    -1.53
    »¿
    -1.51
    scribed
    -1.49
    umm
    -1.46
     bust
    -1.44
    its
    -1.42
    existent
    -1.39
    ubern
    -1.36
    cludes
    -1.35
    POSITIVE LOGITS
     gloves
    1.47
     arrangements
    1.41
    ]{.
    1.36
    anne
    1.35
     Zar
    1.33
     standards
    1.32
    [^
    1.28
    **:
    1.28
    metrics
    1.25
     (\<
    1.23
    Activations Density 1.608%

    No Known Activations