© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 2-RES-SM
    11. 28402
    Prev
    Next
    INDEX
    Explanations

    terms related to uncertainty or lack of information

    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
    326
    +0.11
    0.6%
    488
    +0.11
    0.6%
    334
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    268
    +0.11
    0.02
    277
    +0.11
    0.02
    334
    +0.10
    0.02
    Negative Logits
    ups
    -1.89
    heet
    -1.73
    boards
    -1.62
    ('',
    -1.60
    orent
    -1.52
    up
    -1.48
    heets
    -1.48
    kte
    -1.47
    anner
    -1.40
    otherapy
    -1.40
    POSITIVE LOGITS
     unknown
    1.98
    ness
    1.87
     reasons
    1.84
     whereabouts
    1.76
    liness
    1.73
     intentions
    1.64
     guilt
    1.63
    idades
    1.61
     motives
    1.60
     dates
    1.60
    Activations Density 0.144%

    No Known Activations