© 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. 9705
    Prev
    Next
    INDEX
    Explanations

    keywords related to document structure and attributes

    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
    78
    +0.17
    1.1%
    271
    +0.12
    0.8%
    392
    +0.10
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    78
    +0.17
    -0.00
    92
    +0.12
    0.01
    255
    +0.10
    -0.00
    Negative Logits
     ribbon
    -1.43
     overload
    -1.39
     any
    -1.37
     RSS
    -1.36
     silence
    -1.31
    naments
    -1.27
     defaults
    -1.26
     urge
    -1.25
     favorites
    -1.24
     count
    -1.21
    POSITIVE LOGITS
    ©
    2.88
    Ĩ
    2.79
    Į
    2.62
    ħ
    2.60
    Ļ
    2.55
    ĩ
    2.48
    ĵ
    2.45
    į
    2.42
    ij
    2.39
    ¤
    2.34
    Activations Density 0.041%

    No Known Activations