© 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. 30250
    Prev
    Next
    INDEX
    Explanations

    words related to persistence or remaining in a situation

    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
    115
    +0.13
    0.7%
    59
    +0.12
    0.7%
    77
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    115
    +0.13
    0.02
    59
    +0.12
    0.01
    474
    +0.10
    0.02
    Negative Logits
    ħ
    -2.25
    ¶
    -2.23
    Ī
    -2.11
    Ĩ
    -2.02
    -1.99
     ↵
    -1.99
    ↵
    -1.99
             ↵
    -1.99
    -1.99
    -1.99
    POSITIVE LOGITS
    lers
    1.81
     below
    1.78
    ling
    1.70
    ely
    1.69
    ler
    1.68
    warz
    1.68
    holder
    1.67
    erl
    1.63
    eling
    1.60
    lings
    1.56
    Activations Density 0.163%

    No Known Activations