© 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. 11541
    Prev
    Next
    INDEX
    Explanations

    phrases related to recovery from various types of injuries or trauma

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "recovering" and related forms indicating recovery or improvement

    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
    iances
    -0.66
     pav
    -0.64
    Vote
    -0.60
    ibl
    -0.59
     Sign
    -0.59
    arlane
    -0.59
     Tobacco
    -0.58
    style
    -0.58
    TIT
    -0.58
    arna
    -0.58
    POSITIVE LOGITS
     recovering
    3.74
     recovers
    1.78
     recovery
    1.72
     recover
    1.71
     recovered
    1.69
     Recover
    1.61
     reeling
    1.52
     restoring
    1.38
     returning
    1.35
     undergoing
    1.32
    Activations Density 0.019%

    No Known Activations