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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 0-RES-JB
    11. 20280
    Prev
    Next
    INDEX
    Explanations

    words related to institutions or organizations, specifically those involving research or education

    oai_token-act-pair · gpt-3.5-turbo

    references to units of time, specifically related to duration and intervals

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Asia
    -0.62
     Asset
    -0.61
     theme
    -0.60
     bul
    -0.60
     counsel
    -0.58
     Apost
    -0.57
     baked
    -0.57
     freeze
    -0.57
     headline
    -0.57
     bleed
    -0.56
    POSITIVE LOGITS
    utes
    4.63
    ute
    3.11
    uting
    2.48
    UTE
    2.32
    uted
    2.26
    utions
    1.77
    utations
    1.67
    ution
    1.63
    uters
    1.58
    utation
    1.53
    Activations Density 0.005%

    No Known Activations