© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 8724
    Prev
    Next
    INDEX
    Explanations

    references to software licensing terms and conditions

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    280
    +0.15
    0.8%
    240
    +0.14
    0.8%
    21
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    280
    +0.15
    0.03
    59
    +0.14
    0.05
    136
    +0.13
    -0.00
    Negative Logits
    aha
    -1.82
     Ms
    -1.64
     however
    -1.64
     [...]
    -1.33
    ani
    -1.33
     ladies
    -1.31
    noreply
    -1.30
    olutely
    -1.30
    etric
    -1.29
     manner
    -1.28
    POSITIVE LOGITS
    £
    2.63
    »
    2.56
    ·
    2.50
    ª
    2.48
    ¶
    2.47
    IJ
    2.42
    ¢
    2.39
    Ĥ¬
    2.37
    ¡
    2.33
    §
    2.32
    Activations Density 0.212%

    No Known Activations