© 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. 12835
    Prev
    Next
    INDEX
    Explanations

    phrases related to intent or purpose

    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
    466
    +0.17
    1.0%
    78
    +0.14
    0.8%
    320
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    406
    +0.17
    0.04
    320
    +0.14
    0.06
    393
    +0.12
    0.02
    Negative Logits
    ĻĤ
    -3.26
    ¦
    -3.09
    ³
    -3.09
    Ħ
    -3.01
    Ł
    -2.91
    Ģ
    -2.91
    ij
    -2.84
    Ĵ
    -2.82
    ¨
    -2.78
    İ
    -2.76
    POSITIVE LOGITS
     develop
    2.72
     grow
    2.63
     evolve
    2.41
     accumulate
    2.36
     decline
    2.26
     lose
    2.21
     dominate
    2.20
     suffer
    2.17
     degrade
    2.12
     worry
    2.12
    Activations Density 0.298%

    No Known Activations