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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 12182
    Prev
    Next
    INDEX
    Explanations

    references to legal entities and court-related terminology

    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
    111
    +0.42
    2.8%
    156
    +0.40
    2.6%
    17
    +0.18
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    156
    +0.42
    0.04
    4
    +0.40
    0.04
    111
    +0.18
    0.03
    Negative Logits
    ¿½
    -3.21
    ħ
    -2.87
    ĻĤ
    -2.79
    ¾
    -2.56
    ®
    -2.52
    ¦
    -2.50
    Ľ
    -2.46
    µ
    -2.45
    ²
    -2.44
    º
    -2.41
    POSITIVE LOGITS
    anks
    1.47
    ensive
    1.46
    ract
    1.42
    ender
    1.40
    esters
    1.39
    endra
    1.39
    agra
    1.35
    ough
    1.34
    alent
    1.33
    ensibly
    1.32
    Activations Density 0.040%

    No Known Activations