Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 29149
    Prev
    Next
    INDEX
    Explanations

    terms related to combustion and burning processes

    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
    156
    +0.18
    1.1%
    37
    +0.18
    1.0%
    47
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    37
    +0.18
    0.01
    389
    +0.18
    0.03
    115
    +0.15
    0.02
    Negative Logits
    ···
    -1.92
     vaccinated
    -1.76
     donated
    -1.73
     married
    -1.59
    amente
    -1.58
    reon
    -1.53
     gestational
    -1.50
     postnatal
    -1.47
    clerosis
    -1.47
     nesting
    -1.47
    POSITIVE LOGITS
    Ļª
    2.54
    iston
    2.15
     combustion
    1.94
     fires
    1.83
    stock
    1.81
    Ĵ
    1.78
     gases
    1.78
    Ģ
    1.76
    ·
    1.73
    gas
    1.73
    Activations Density 0.222%

    No Known Activations