© 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. 4-RES-SM
    11. 9477
    Prev
    Next
    INDEX
    Explanations

    proper names and specific identifiers

    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/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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.19
    1.1%
    466
    +0.15
    0.9%
    410
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    410
    +0.19
    0.13
    266
    +0.15
    0.11
    477
    +0.12
    0.12
    Negative Logits
    ģ
    -3.85
    ↵  ↵
    -3.84
    -3.84
    <|padding|>
    -3.84
      ↵
    -3.84
    <|outofrange|>
    -3.84
    ↵
    -3.84
    ↵
    -3.84
    ↵  âĢĥ
    -3.84
    <|outofrange|>
    -3.84
    POSITIVE LOGITS
    ovic
    2.54
    udeau
    2.44
    iewicz
    2.44
    hof
    2.37
    owski
    2.35
    ovich
    2.32
    hoff
    2.31
    emann
    2.29
    opoulos
    2.27
    hardt
    2.25
    Activations Density 0.832%

    No Known Activations