© 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. 4742
    Prev
    Next
    INDEX
    Explanations

    terms related to motion or movement

    oai_token-act-pair · gpt-3.5-turbo

    references to motion pictures and related terminology

    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
     trough
    -0.80
     hole
    -0.78
     peg
    -0.74
     begg
    -0.71
     communal
    -0.71
     absorb
    -0.68
     symbol
    -0.67
     estab
    -0.66
     hook
    -0.66
     hoard
    -0.66
    POSITIVE LOGITS
     Motion
    3.35
    Motion
    2.29
    motion
    1.80
     MOT
    1.31
     Speed
    1.15
     Ventures
    1.11
     Recording
    1.09
     Performance
    1.05
    Performance
    1.00
     Notification
    1.00
    Activations Density 0.051%

    No Known Activations