© 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. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 32906
    Prev
    Next
    INDEX
    Explanations

    phrases related to potential actions or future possibilities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     standards
    -0.21
     Standards
    -0.20
     standardized
    -0.17
    rschein
    -0.16
    .synthetic
    -0.15
    æłĩåĩĨ
    -0.15
     Architects
    -0.15
    缮ãĤĴ
    -0.15
     standard
    -0.14
     streets
    -0.14
    POSITIVE LOGITS
     sentinel
    0.19
     pillar
    0.17
     apart
    0.17
    ISTA
    0.16
     ?>"/>↵
    0.16
     attention
    0.16
    reason
    0.16
    íĴĪ
    0.15
    alone
    0.15
    ismet
    0.15
    Activations Density 0.005%

    No Known Activations