{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/d2po-discriminator-guided-dpo-with-response","title":"D2PO: Discriminator-Guided DPO with Response Evaluation Models","arxiv_id":"2405.01511","date":"2024-05-02","proceeding":null,"authors":["Prasann Singhal","Nathan Lambert","Scott Niekum","Tanya Goyal","Greg Durrett"],"abstract":"Varied approaches for aligning language models have been proposed, including supervised fine-tuning, RLHF, and direct optimization methods such as DPO. Although DPO has rapidly gained popularity due to its straightforward training process and competitive results, there is an open question of whether there remain practical advantages of using a discriminator, like a reward model, to evaluate responses. We propose D2PO, discriminator-guided DPO, an approach for the online setting where preferences are being collected throughout learning. As we collect gold preferences, we use these not only to train our policy, but to train a discriminative response evaluation model to silver-label even more synthetic data for policy training. We explore this approach across a set of diverse tasks, including a realistic chat setting, we find that our approach leads to higher-quality outputs compared to DPO with the same data budget, and greater efficiency in terms of preference data requirements. Furthermore, we show conditions under which silver labeling is most helpful: it is most effective when training the policy with DPO, outperforming traditional PPO, and benefits from maintaining a separate discriminator from the policy model.","url_abs":"https://arxiv.org/abs/2405.01511v2","url_pdf":"https://arxiv.org/pdf/2405.01511v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"d2po-discriminator-guided-dpo-with-response","repo_url":"https://github.com/PrasannS/d2po","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[],"methods":[{"method_slug":"dpo","method_name":"DPO"},{"method_slug":"entropy-regularization","method_name":"Entropy Regularization"},{"method_slug":"ppo","method_name":"PPO"},{"method_slug":"set","method_name":"SET"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2405.01511","atlas_url":"https://app.syntology.ai/?focus=2405.01511","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2405.01511"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/PrasannS/d2po","reach":{"status":"ok"}}],"summary":{"ran":11,"unverified":2},"by_repo_kind":{"official":{"samples":13,"ran":11,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":13,"samples":[{"code_sha256_prefix":"bdf1e0f893679196","entry":"add_row_index","repo":"PrasannS/d2po","repo_kind":"official","path":"src/train_rm.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/train_rm.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"bdf1e0f893679196"}},{"code_sha256_prefix":"39dae513044dc499","entry":"adjust_input","repo":"PrasannS/d2po","repo_kind":"official","path":"src/eval/generate_outs.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/eval/generate_outs.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"39dae513044dc499"}},{"code_sha256_prefix":"22de807813131078","entry":"adjust_tulu","repo":"PrasannS/d2po","repo_kind":"official","path":"src/eval/generate_outs.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/eval/generate_outs.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"22de807813131078"}},{"code_sha256_prefix":"af95fc70de192cc3","entry":"get_step_ckpt","repo":"PrasannS/d2po","repo_kind":"official","path":"src/eval/generate_outs.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/eval/generate_outs.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"af95fc70de192cc3"}},{"code_sha256_prefix":"ea8d7cf234eea3bb","entry":"lensco","repo":"PrasannS/d2po","repo_kind":"official","path":"src/utils/rl_utils.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/utils/rl_utils.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ea8d7cf234eea3bb"}},{"code_sha256_prefix":"d0004fe553444d0e","entry":"new_forward","repo":"PrasannS/d2po","repo_kind":"official","path":"rlhfutils/rlhfutils/modeling_override.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/rlhfutils/rlhfutils/modeling_override.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d0004fe553444d0e"}},{"code_sha256_prefix":"80a8b3f87c775a21","entry":"onlyans","repo":"PrasannS/d2po","repo_kind":"official","path":"src/train_dpo.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/train_dpo.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"80a8b3f87c775a21"}},{"code_sha256_prefix":"148dc2c38b7cd922","entry":"prepare_sample_text","repo":"PrasannS/d2po","repo_kind":"official","path":"src/trainsft.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/trainsft.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"148dc2c38b7cd922"}},{"code_sha256_prefix":"3f0252eb62a38d86","entry":"preproc_wgpt","repo":"PrasannS/d2po","repo_kind":"official","path":"src/eval/generate_outs_old.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/eval/generate_outs_old.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3f0252eb62a38d86"}},{"code_sha256_prefix":"b4cda67a5a5317e9","entry":"simplecat","repo":"PrasannS/d2po","repo_kind":"official","path":"src/train_dpo.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/train_dpo.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b4cda67a5a5317e9"}},{"code_sha256_prefix":"ab9274275c75ba41","entry":"tulu_pf","repo":"PrasannS/d2po","repo_kind":"official","path":"src/train_dpo.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/train_dpo.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"ab9274275c75ba41"}},{"code_sha256_prefix":"576b5052f87d577b","entry":"get_pipeline","repo":"PrasannS/d2po","repo_kind":"official","path":"src/utils/rl_utils.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/utils/rl_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"576b5052f87d577b"}},{"code_sha256_prefix":"8faae6d33f36ed52","entry":"load_stack","repo":"PrasannS/d2po","repo_kind":"official","path":"src/eval/generate_outs_old.py","file_url":"https://github.com/PrasannS/d2po/blob/HEAD/src/eval/generate_outs_old.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8faae6d33f36ed52"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}