© 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. 14221
    Prev
    Next
    INDEX
    Explanations

    references to folders and directory structures

    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
    156
    +0.22
    1.3%
    171
    +0.13
    0.7%
    217
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.22
    0.00
    171
    +0.13
    0.01
    172
    +0.11
    0.01
    Negative Logits
     deaf
    -1.73
    ureus
    -1.69
    erad
    -1.67
     appetite
    -1.55
    ity
    -1.50
     voices
    -1.47
     speaking
    -1.45
    lang
    -1.41
     breath
    -1.40
     quo
    -1.39
    POSITIVE LOGITS
    Ģ
    1.89
    Ĺ
    1.74
    ĩ
    1.69
    ĻĤ
    1.66
    ĭ
    1.57
    holders
    1.57
    ocs
    1.55
     possessions
    1.51
    ij
    1.51
     size
    1.48
    Activations Density 0.128%

    No Known Activations