© 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. 7-RES-JB
    11. 22625
    Prev
    Next
    INDEX
    Explanations

    the word "hence" and phrases related to cause and effect

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "hence" indicating causal relationships or logical conclusions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     convey
    -0.60
     rehabilit
    -0.60
     crochet
    -0.59
     batter
    -0.59
    estation
    -0.58
    CR
    -0.58
     hitter
    -0.55
     Derby
    -0.55
     Killing
    -0.55
     battered
    -0.54
    POSITIVE LOGITS
    forth
    2.07
    forward
    1.41
     why
    0.93
    noon
    0.88
    fter
    0.85
    ãĤ´ãĥ³
    0.84
    pend
    0.83
    xual
    0.80
    alf
    0.79
    rely
    0.79
    Activations Density 0.020%

    No Known Activations