© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. MLP Post
    9. 0-MLP-SM
    11. 29085
    Prev
    Next
    INDEX
    Explanations

    the adjective "pretty" used to describe various subjects

    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
    376
    +0.18
    1.0%
    444
    +0.12
    0.7%
    255
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    374
    +0.18
    0.04
    255
    +0.12
    0.04
    262
    +0.11
    0.04
    Negative Logits
     BGCOLOR
    -1.55
    ^*
    -1.53
     time
    -1.52
    xture
    -1.45
     only
    -1.44
     grade
    -1.42
     answer
    -1.39
     present
    -1.34
     opportunity
    -1.32
     disability
    -1.30
    POSITIVE LOGITS
    ¿½
    3.38
    ĻĤ
    3.07
    Īĺ
    2.85
    ³
    2.62
            ↵
    2.62
    <|outofrange|>
    2.62
    2.62
    ↵↵↵
    2.62
    ↵↵
    2.62
    ↵↵
    2.62
    Activations Density 0.031%

    No Known Activations