Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385890247> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4385890247 abstract "Chain-of-Thought (CoT) prompting in large language models (LLMs) has shown promising performance on mathematical reasoning tasks. Recently, Self-Consistency samples a diverse set of reasoning chains with different answers and chooses the answer by majority voting. Though effective, its performance cannot be further improved by sampling more reasoning chains. To address this problem, we propose to integrate backward reasoning into answer verification. We first mask a number in the question by ${bf x}$. The LLM is then asked to predict the masked number with a candidate answer $A$ embedded in the template: ``If we know the answer to the above question is ${A}$, what is the value of unknown variable ${bf x}$?'' The LLM is expected to predict the masked number successfully if the provided candidate answer is correct. To further improve performance, we propose FOBAR (FOrward-BAckward Reasoning) to combine forward and backward reasoning for verifying candidate answers. Experiments are performed on six standard mathematical data sets and three LLMs (text-davinci-003, GPT-3.5-Turbo, GPT-4). Results show that FOBAR achieves state-of-the-art performance. In particular, FOBAR outperforms Self-Consistency which uses forward reasoning alone, demonstrating that combining forward and forward reasoning is better. It also outperforms existing verification methods, verifying the effectiveness of using the simple template in backward reasoning and the proposed combination." @default.
- W4385890247 created "2023-08-17" @default.
- W4385890247 creator A5026487799 @default.
- W4385890247 creator A5033066610 @default.
- W4385890247 creator A5049419584 @default.
- W4385890247 creator A5066103173 @default.
- W4385890247 creator A5070273088 @default.
- W4385890247 creator A5071773009 @default.
- W4385890247 creator A5077862962 @default.
- W4385890247 date "2023-08-15" @default.
- W4385890247 modified "2023-10-03" @default.
- W4385890247 title "Forward-Backward Reasoning in Large Language Models for Mathematical Verification" @default.
- W4385890247 doi "https://doi.org/10.48550/arxiv.2308.07758" @default.
- W4385890247 hasPublicationYear "2023" @default.
- W4385890247 type Work @default.
- W4385890247 citedByCount "0" @default.
- W4385890247 crossrefType "posted-content" @default.
- W4385890247 hasAuthorship W4385890247A5026487799 @default.
- W4385890247 hasAuthorship W4385890247A5033066610 @default.
- W4385890247 hasAuthorship W4385890247A5049419584 @default.
- W4385890247 hasAuthorship W4385890247A5066103173 @default.
- W4385890247 hasAuthorship W4385890247A5070273088 @default.
- W4385890247 hasAuthorship W4385890247A5071773009 @default.
- W4385890247 hasAuthorship W4385890247A5077862962 @default.
- W4385890247 hasBestOaLocation W43858902471 @default.
- W4385890247 hasConcept C119857082 @default.
- W4385890247 hasConcept C154945302 @default.
- W4385890247 hasConcept C177264268 @default.
- W4385890247 hasConcept C195344581 @default.
- W4385890247 hasConcept C199360897 @default.
- W4385890247 hasConcept C204321447 @default.
- W4385890247 hasConcept C2776436953 @default.
- W4385890247 hasConcept C41008148 @default.
- W4385890247 hasConceptScore W4385890247C119857082 @default.
- W4385890247 hasConceptScore W4385890247C154945302 @default.
- W4385890247 hasConceptScore W4385890247C177264268 @default.
- W4385890247 hasConceptScore W4385890247C195344581 @default.
- W4385890247 hasConceptScore W4385890247C199360897 @default.
- W4385890247 hasConceptScore W4385890247C204321447 @default.
- W4385890247 hasConceptScore W4385890247C2776436953 @default.
- W4385890247 hasConceptScore W4385890247C41008148 @default.
- W4385890247 hasLocation W43858902471 @default.
- W4385890247 hasOpenAccess W4385890247 @default.
- W4385890247 hasPrimaryLocation W43858902471 @default.
- W4385890247 hasRelatedWork W2353865532 @default.
- W4385890247 hasRelatedWork W2961085424 @default.
- W4385890247 hasRelatedWork W3046775127 @default.
- W4385890247 hasRelatedWork W3170094116 @default.
- W4385890247 hasRelatedWork W4205958290 @default.
- W4385890247 hasRelatedWork W4285260836 @default.
- W4385890247 hasRelatedWork W4286629047 @default.
- W4385890247 hasRelatedWork W4306321456 @default.
- W4385890247 hasRelatedWork W4306674287 @default.
- W4385890247 hasRelatedWork W4224009465 @default.
- W4385890247 isParatext "false" @default.
- W4385890247 isRetracted "false" @default.
- W4385890247 workType "article" @default.