© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 4-RES-SM
    6. 12185
    Prev
    Next
    INDEX
    Explanations

    concepts related to inventions and disclosures

    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
    21
    +0.27
    1.7%
    78
    +0.14
    0.8%
    63
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    334
    +0.27
    0.01
    63
    +0.14
    0.00
    216
    +0.10
    0.01
    Negative Logits
    autres
    -1.42
     attendance
    -1.35
     others
    -1.30
    lining
    -1.30
    OOGLE
    -1.30
    cue
    -1.29
    otin
    -1.29
    hell
    -1.26
     seriously
    -1.25
    others
    -1.22
    POSITIVE LOGITS
     relates
    1.83
     pertains
    1.73
     disclosed
    1.60
     herein
    1.58
    ursuant
    1.54
     discloses
    1.51
     concerns
    1.51
     generally
    1.50
     deals
    1.48
     comprises
    1.44
    Activations Density 0.006%

    No Known Activations