© 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 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. 11054
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly periods

    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
    +0.45
    3.0%
    111
    +0.39
    2.6%
    271
    +0.13
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.45
    0.44
    156
    +0.39
    0.31
    71
    +0.13
    0.31
    Negative Logits
      ↵
    -2.99
    <|padding|>
    -2.99
    ↵
    -2.99
    -2.99
    <|outofrange|>
    -2.99
    č↵
    -2.99
    ↵↵
    -2.99
    ↵
    -2.99
    ↵↵
    -2.99
    ↵
    -2.99
    POSITIVE LOGITS
    ension
    1.65
    wick
    1.62
    arget
    1.47
    assen
    1.45
    ungen
    1.45
    uth
    1.43
    emp
    1.43
    acker
    1.41
    ublin
    1.40
    pendicular
    1.38
    Activations Density 0.438%

    No Known Activations