PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking
Paper
•
2410.12375
•
Published
•
2
PRefLexOR
PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking
Note Paper on arXiv
lamm-mit/PRefLexOR_ORPO_DPO_EXO_10242024
Text Generation • Updated • 69Note Model produces thinking tokens before answering
lamm-mit/PRefLexOR_ORPO_DPO_EXO_REFLECT_10222024
Text Generation • Updated • 20 • 1Note Model produces both thinking and reflection tokens before answering
lamm-mit/meta-llama-Meta-Llama-3.2-3B-Instruct-Reasoning-Tokenizer
UpdatedNote Llama tokenizer with special tokens added (thinking, reflection, scratchpad, response)