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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 8-RES-JB
    6. 21171
    Prev
    Next
    INDEX
    Explanations

    mentions of different types of crimes

    oai_token-act-pair · gpt-3.5-turbo

    references to crimes, particularly those against humanity

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    quickShipAvailable
    -0.81
    soDeliveryDate
    -0.70
    atari
    -0.70
    alg
    -0.69
    por
    -0.68
    visors
    -0.67
    flush
    -0.67
    pole
    -0.66
     snipp
    -0.64
    achev
    -0.64
    POSITIVE LOGITS
     punishable
    1.40
     committed
    1.22
     perpetrated
    1.14
     offences
    1.07
     spree
    1.00
     crimes
    0.97
     offenses
    0.96
     indict
    0.95
     convictions
    0.95
     misdem
    0.92
    Activations Density 0.079%

    No Known Activations