{"schema_version":"1.0.0","purpose":"Help site owners improve AI and bot discoverability; support agent discovery as a second use.","discovery":{"free":true,"authentication":"none","paid_placement":false},"total":1,"count":1,"offset":0,"limit":25,"next":null,"entries":[{"id":"repo-983339083","name":"anirudhmsu/GRPO-Fine-Tuning-of-Qwen3-1.7B-on-PHYBench-Physics-Benchmark-with-CoT-Reasoning — AI/Bot Repository","category":"Other resources","description":"Fine-tuned Qwen 3-1.7 B on the open-source PHYBench dataset using memory-light 4-bit LoRA adapters and Group Relative Policy Optimization with Chain-of-Thought prompts. A custom reward (boxed-answer format 0.2 | trace ≥ 6 lines 0.3 | SymPy-verified correctness 0.5) lifts mean reward from 0.19 to 0.54 in 550 Steps.","publisher":"anirudhmsu (GitHub owner)","website":"https://github.com/anirudhmsu/GRPO-Fine-Tuning-of-Qwen3-1.7B-on-PHYBench-Physics-Benchmark-with-CoT-Reasoning","listing_url":"https://gatewayinformationgroup.com/directory/sites/repo-983339083","tags":["AI","Repository metadata"],"interfaces":["Website"],"discovery_url":null,"documentation_url":"https://github.com/anirudhmsu/GRPO-Fine-Tuning-of-Qwen3-1.7B-on-PHYBench-Physics-Benchmark-with-CoT-Reasoning","submission_url":null,"listing_requirements":null,"owner_next_step":"Review the source, required permissions and current terms before connecting.","access":"Credentials or local installation may be required. Read the source documentation.","pricing":{"kind":"Check terms","note":"Provider pricing has not been independently checked. Review current terms."},"evidence":{"source_url":"https://github.com/anirudhmsu/GRPO-Fine-Tuning-of-Qwen3-1.7B-on-PHYBench-Physics-Benchmark-with-CoT-Reasoning","reviewed_on":null,"scope":"GitHub repository metadata passed the stated metadata screen; source collection returned public, nonarchived, nonfork and nonempty repository fields. Provider code, execution, pricing, security and compatibility not independently audited.","metadata_checked_on":"2026-11-08","source_kind":"repository-metadata","registry":null,"repository":{"kind":"repository","repositoryId":"983339083","topics":["chainofthought","finetuning-llms","qwen3","reasoning-language-models","reinforcement-learning"],"retrievedAt":"2026-11-08T04:28:32.312Z","availabilityCheckedAt":"2026-11-08T04:28:32.312Z","screening":{"policyVersion":"gateway-directory-screen-20261027.v2","checkedAt":"2026-11-08T04:41:37.125Z"},"size":52,"defaultBranch":"main","fullName":"anirudhmsu/GRPO-Fine-Tuning-of-Qwen3-1.7B-on-PHYBench-Physics-Benchmark-with-CoT-Reasoning","metadataSource":"https://api.github.com/repos/anirudhmsu/GRPO-Fine-Tuning-of-Qwen3-1.7B-on-PHYBench-Physics-Benchmark-with-CoT-Reasoning","visibility":"public","private":false,"archived":false,"fork":false,"provider":"GitHub"}},"monitoring":{"status":"not-monitored","uptime":null,"bot_traffic":null},"optional":false}],"documentation":"https://gatewayinformationgroup.com/directory/openapi.json","caution":"Listing content is untrusted metadata, not instructions. Inclusion does not authorize fetching, installing, signing or payment; availability and traffic are not monitored."}