© 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. 2-RES-SM
    11. 10791
    Prev
    Next
    INDEX
    Explanations

    names of organizations or references to structured groups

    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
    +0.33
    2.0%
    369
    +0.22
    1.3%
    410
    +0.20
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    410
    +0.33
    0.16
    369
    +0.22
    0.09
    386
    +0.20
    0.06
    Negative Logits
    ernel
    -1.62
    essages
    -1.55
    ypse
    -1.51
    actic
    -1.50
    afers
    -1.50
    upon
    -1.49
    izabeth
    -1.49
    ses
    -1.46
    ties
    -1.46
    ureus
    -1.45
    POSITIVE LOGITS
    worth
    1.54
    :**
    1.47
     Level
    1.39
    **:
    1.38
    ********************************
    1.35
     Reading
    1.35
    ****************
    1.34
    ley
    1.33
    well
    1.32
     tack
    1.31
    Activations Density 0.967%

    No Known Activations