© 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 EvalsExportsinterp-engineNEWAPI 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. 13805
    Prev
    Next
    INDEX
    Explanations

    phrases and references to various areas or topics

    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
    98
    +0.17
    1.0%
    419
    +0.14
    0.8%
    161
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    16
    +0.17
    0.04
    457
    +0.14
    0.04
    95
    +0.14
    0.04
    Negative Logits
     drums
    -1.57
    ickets
    -1.52
    auf
    -1.50
     torch
    -1.49
    pler
    -1.48
    rons
    -1.38
     drum
    -1.36
     directions
    -1.33
     burgers
    -1.32
    ÑĮ
    -1.31
    POSITIVE LOGITS
    »
    2.66
    ĵ
    2.41
    Ń
    2.25
    »¿
    2.11
    ·¸
    2.07
    ¼
    2.03
    Ħ
    2.01
    ¬
    1.99
    ´
    1.97
    ł
    1.95
    Activations Density 0.275%

    No Known Activations