© 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. 2-RES-SM
    11. 32394
    Prev
    Next
    INDEX
    Explanations

    references to software version numbers

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    320
    +0.12
    0.7%
    358
    +0.12
    0.7%
    230
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    329
    +0.12
    0.02
    222
    +0.12
    0.01
    230
    +0.10
    0.01
    Negative Logits
    ?_
    -1.86
    ._
    -1.79
    _."
    -1.66
    ."_
    -1.66
    Īĺ
    -1.63
    Ĺ
    -1.59
    "?"
    -1.55
    ·¸
    -1.53
    ļ
    -1.52
    ¶
    -1.50
    POSITIVE LOGITS
     authored
    1.48
     waiver
    1.45
    ÑĮ
    1.44
     guarantee
    1.41
    dk
    1.41
     discretion
    1.40
     Clause
    1.38
     guarantees
    1.36
    forge
    1.35
    License
    1.30
    Activations Density 0.033%

    No Known Activations