© 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. 2-RES-SM
    11. 31805
    Prev
    Next
    INDEX
    Explanations

    keywords related to filtering processes and conditions

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    +0.20
    1.2%
    354
    +0.11
    0.7%
    230
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    230
    +0.20
    0.01
    173
    +0.11
    0.01
    443
    +0.11
    0.01
    Negative Logits
    wards
    -1.96
    ±
    -1.95
    ĥ½
    -1.64
    usement
    -1.57
    lla
    -1.57
    ---|---
    -1.57
    unnumbered
    -1.51
     HREF
    -1.47
    terday
    -1.45
    ½
    -1.43
    POSITIVE LOGITS
    ings
    1.75
     enforcement
    1.61
    suits
    1.57
    ament
    1.57
    isher
    1.54
    codes
    1.53
    caster
    1.52
    ership
    1.52
    Chain
    1.51
    eries
    1.44
    Activations Density 0.123%

    No Known Activations