© 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. MLP Post
    5. 0-MLP-SM
    6. 27981
    Prev
    Next
    INDEX
    Explanations

    references to aluminum

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__mlp-sm_processed/0-mlp-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.0.hook_mlp_out
    Hook Layer
    0
    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
    +0.23
    1.3%
    443
    +0.12
    0.7%
    294
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    69
    +0.23
    0.04
    434
    +0.12
    0.03
    97
    +0.11
    0.02
    Negative Logits
    rapy
    -1.73
    ATH
    -1.39
     solitary
    -1.32
    ÑĢоÑģ
    -1.30
    itoneal
    -1.30
    ellation
    -1.28
    iÄĩ
    -1.28
    atory
    -1.27
    ages
    -1.26
    athered
    -1.26
    POSITIVE LOGITS
    º
    1.91
    erals
    1.81
    ppo
    1.79
    hower
    1.75
    ¹
    1.72
    ģ
    1.72
    ieu
    1.67
    ¶
    1.67
    ²
    1.67
    illa
    1.56
    Activations Density 1.712%

    No Known Activations