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. 12100
    Prev
    Next
    INDEX
    Explanations

    legal terms and phrases related to imprisonment and sentencing

    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
    290
    +0.14
    0.8%
    394
    +0.13
    0.8%
    251
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    177
    +0.14
    0.04
    251
    +0.13
    0.04
    359
    +0.12
    0.04
    Negative Logits
    ı
    -3.01
    Ĵ
    -2.75
    ĭ
    -2.75
    ĵ
    -2.73
    ĥ½
    -2.68
    Īĺ
    -2.64
    ¨
    -2.57
    ®
    -2.54
    ¹
    -2.52
      ↵   
    -2.50
    POSITIVE LOGITS
     eligibility
    1.80
     viol
    1.76
     amount
    1.76
     procedure
    1.75
     inmate
    1.74
    break
    1.72
     violations
    1.71
     offenders
    1.70
    itude
    1.66
     sentence
    1.65
    Activations Density 0.600%

    No Known Activations