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

    Neuronpedia

    Jacobian LensNEW
    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. 3-RES-JB
    11. 5222
    Prev
    Next
    INDEX
    Explanations

    phrases related to giving or offering something

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to giving or distributing items

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ģ«
    -0.69
     thous
    -0.67
     Rhode
    -0.67
    nel
    -0.66
     currents
    -0.64
     Presbyter
    -0.62
     attrition
    -0.61
     Modes
    -0.60
    atos
    -0.60
    issy
    -0.60
    POSITIVE LOGITS
     amnesty
    0.73
     chance
    0.66
     airs
    0.66
    free
    0.66
     berth
    0.65
     pause
    0.65
     glimpse
    0.64
    summary
    0.63
    license
    0.63
     ample
    0.63
    Activations Density 0.220%

    No Known Activations