© 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. 24891
    Prev
    Next
    INDEX
    Explanations

    references to sessions in various 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
    +0.24
    1.4%
    361
    +0.11
    0.6%
    500
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    498
    +0.24
    0.02
    184
    +0.11
    0.02
    10
    +0.11
    0.01
    Negative Logits
     sexually
    -1.65
     MERCHANTABILITY
    -1.63
     racial
    -1.52
    iquity
    -1.50
    sexual
    -1.42
     way
    -1.42
    Ĵ
    -1.42
     eggs
    -1.39
     Catholic
    -1.38
    oxin
    -1.36
    POSITIVE LOGITS
     duration
    1.72
     durations
    1.72
    now
    1.65
    controller
    1.63
     timer
    1.57
    wright
    1.57
     controller
    1.51
    etable
    1.47
    ive
    1.47
    table
    1.46
    Activations Density 0.144%

    No Known Activations