© 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. 12527
    Prev
    Next
    INDEX
    Explanations

    references to chemical compounds and reactions

    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
    2
    +0.17
    0.9%
    287
    +0.14
    0.8%
    448
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    2
    +0.17
    0.02
    360
    +0.14
    0.02
    287
    +0.12
    0.02
    Negative Logits
    chitz
    -2.24
    ŀ
    -2.21
    ľĵ
    -2.01
    Ĺ
    -2.01
    avirus
    -1.95
    UAL
    -1.91
    ij
    -1.90
    ĥ½
    -1.89
    Ļª
    -1.85
    ureus
    -1.82
    POSITIVE LOGITS
     composition
    2.24
     compounds
    2.20
     synthesis
    2.12
     enrichment
    2.10
     substances
    2.06
     additives
    2.05
     mixtures
    2.01
     compositions
    1.97
     reagents
    1.97
     vapor
    1.92
    Activations Density 0.048%

    No Known Activations