© 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. 1-RES-SM
    11. 29322
    Prev
    Next
    INDEX
    Explanations

    references to specific types of gels, particularly in scientific contexts

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    +0.17
    1.0%
    10
    +0.13
    0.8%
    386
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    386
    +0.17
    0.01
    43
    +0.13
    0.01
    115
    +0.11
    0.01
    Negative Logits
    woke
    -1.86
     happier
    -1.82
     retiring
    -1.76
     slightest
    -1.69
     say
    -1.66
     murm
    -1.66
     victims
    -1.64
     gay
    -1.59
    TRODUCTION
    -1.57
     hesitate
    -1.53
    POSITIVE LOGITS
    dings
    2.02
    ings
    1.79
    ator
    1.77
    atto
    1.66
    anyl
    1.66
    bridge
    1.64
    ister
    1.64
    ded
    1.62
    ambers
    1.59
    ains
    1.59
    Activations Density 0.070%

    No Known Activations