© 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. 8-RES-JB
    11. 19704
    Prev
    Next
    INDEX
    Explanations

    specific identification numbers and related documents needed for various bureaucratic processes

    oai_token-act-pair · gpt-3.5-turbo

    requirements related to identification and documentation processes

    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
    reddits
    -0.73
    gallery
    -0.72
    "]=>
    -0.69
    NetMessage
    -0.68
     speculate
    -0.66
    cffffcc
    -0.66
    RIP
    -0.64
    1027
    -0.64
    erest
    -0.61
    ????
    -0.61
    POSITIVE LOGITS
     authorization
    1.12
     consent
    1.05
     satisfactory
    1.04
     certification
    1.02
     receipt
    0.97
     approval
    0.96
     verification
    0.93
     specified
    0.91
     waiver
    0.91
     affirmative
    0.91
    Activations Density 0.573%

    No Known Activations