© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 11966
    Prev
    Next
    INDEX
    Explanations

    technical jargon and specific numerical data related to scientific topics

    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
    21
    +0.17
    0.9%
    122
    +0.14
    0.8%
    54
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    12
    +0.17
    0.05
    72
    +0.14
    0.05
    356
    +0.12
    0.07
    Negative Logits
    ĨĴ
    -3.03
    ĺ
    -2.77
    Ħ
    -2.63
    ¿½
    -2.51
    §
    -2.37
      ↵
    -2.35
    -2.34
    ↵
    -2.34
    ↵
    -2.34
    č↵
    -2.34
    POSITIVE LOGITS
    ugin
    1.60
     were
    1.48
     was
    1.44
     premiered
    1.43
     reacts
    1.41
     reacted
    1.41
    derived
    1.40
     participated
    1.40
    yield
    1.38
     represents
    1.38
    Activations Density 0.598%

    No Known Activations