© 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. 1-RES-SM
    11. 25194
    Prev
    Next
    INDEX
    Explanations

    possessive pronouns referring to ownership or affiliation

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    23
    +0.30
    1.7%
    478
    +0.15
    0.9%
    369
    +0.15
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    338
    +0.30
    0.06
    240
    +0.15
    0.02
    284
    +0.15
    0.05
    Negative Logits
    ppo
    -1.62
     arena
    -1.58
    zone
    -1.57
    ophone
    -1.56
    amento
    -1.54
     genre
    -1.52
     alc
    -1.51
    rails
    -1.50
     cinema
    -1.48
    acs
    -1.48
    POSITIVE LOGITS
    Ī
    2.79
    ¿½
    2.78
    ķ
    2.65
    IJ
    2.53
    ĵ
    2.51
    Ń
    2.50
    «
    2.50
    ²
    2.44
    ĺ
    2.44
    ¨
    2.43
    Activations Density 0.585%

    No Known Activations