© 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. 4-RES-SM
    11. 12295
    Prev
    Next
    INDEX
    Explanations

    mentions of the term "Mercy" 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/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
    447
    +0.13
    0.7%
    21
    +0.11
    0.6%
    203
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.13
    0.01
    59
    +0.11
    0.02
    203
    +0.11
    0.02
    Negative Logits
    ers
    -2.15
     trouble
    -1.44
     probability
    -1.43
     instant
    -1.43
     mismatch
    -1.40
    emic
    -1.39
    tics
    -1.33
     partir
    -1.33
     measurement
    -1.31
    etry
    -1.30
    POSITIVE LOGITS
    ī
    4.08
    ª
    3.87
    «
    3.71
    ³
    3.68
    ½
    3.60
    Ĵ
    3.53
    Ħ
    3.53
    ¦
    3.53
    į
    3.49
    ¡
    3.49
    Activations Density 0.133%

    No Known Activations