© 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
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 8-RES-JB
    11. 7127
    Prev
    Next
    INDEX
    Explanations

    numeric values or quantities mentioned in text

    oai_token-act-pair · gpt-3.5-turbo

    quantitative references related to prosecutions, investigations, and legislative actions

    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
    nature
    -0.70
    someone
    -0.69
    tis
    -0.67
     Feeling
    -0.67
    ĺħ
    -0.67
     Played
    -0.66
    emi
    -0.65
     Ruler
    -0.63
    oyer
    -0.62
    Redditor
    -0.62
    POSITIVE LOGITS
     preliminary
    0.91
     proposals
    0.87
     packages
    0.83
     prototypes
    0.80
     rounds
    0.80
     consultations
    0.78
     variants
    0.77
     strategies
    0.77
     versions
    0.76
     plans
    0.76
    Activations Density 0.386%

    No Known Activations