{"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/unintended-impacts-of-llm-alignment-on-global","title":"Unintended Impacts of LLM Alignment on Global Representation","arxiv_id":"2402.15018","date":"2024-02-22","proceeding":null,"authors":["Michael J. Ryan","William Held","Diyi Yang"],"abstract":"Before being deployed for user-facing applications, developers align Large Language Models (LLMs) to user preferences through a variety of procedures, such as Reinforcement Learning From Human Feedback (RLHF) and Direct Preference Optimization (DPO). Current evaluations of these procedures focus on benchmarks of instruction following, reasoning, and truthfulness. However, human preferences are not universal, and aligning to specific preference sets may have unintended effects. We explore how alignment impacts performance along three axes of global representation: English dialects, multilingualism, and opinions from and about countries worldwide. Our results show that current alignment procedures create disparities between English dialects and global opinions. We find alignment improves capabilities in several languages. We conclude by discussing design decisions that led to these unintended impacts and recommendations for more equitable preference tuning. We make our code and data publicly available on Github.","url_abs":"https://arxiv.org/abs/2402.15018v2","url_pdf":"https://arxiv.org/pdf/2402.15018v2.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":"unintended-impacts-of-llm-alignment-on-global","repo_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"instruction-following","task_name":"Instruction Following"}],"methods":[{"method_slug":"align","method_name":"ALIGN"},{"method_slug":"focus","method_name":"Focus"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2402.15018","atlas_url":"https://app.syntology.ai/?focus=2402.15018","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.15018"}},"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/salt-nlp/unintended-impacts-of-alignment","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":5,"unverified":6},"by_repo_kind":{"official":{"samples":11,"ran":5,"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":0,"samples":[{"code_sha256_prefix":"292ef38a85bc5792","entry":"compute_negative_log_likelihoods","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/ask-reddit-perplexity-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/ask-reddit-perplexity-testing.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"292ef38a85bc5792"}},{"code_sha256_prefix":"d93bb71a63350c87","entry":"compute_negative_log_likelihoods","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/belebele.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/belebele.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d93bb71a63350c87"}},{"code_sha256_prefix":"6a5d0168a52f7d49","entry":"convert_iso_code_to_language_name","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/belebele.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/belebele.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6a5d0168a52f7d49"}},{"code_sha256_prefix":"c904819afc630c4d","entry":"generate_country_prompts","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/reward-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/reward-testing.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c904819afc630c4d"}},{"code_sha256_prefix":"98a3c96a35040079","entry":"should_add_article","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/ask-reddit-perplexity-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/ask-reddit-perplexity-testing.py","link_basis":"first_harvest_node","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"98a3c96a35040079"}},{"code_sha256_prefix":"ea124a9e59aa3d97","entry":"generate_answer","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/md3game.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/md3game.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ea124a9e59aa3d97"}},{"code_sha256_prefix":"c0b8cf9d3d7e3400","entry":"get_rewards","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/ask-reddit-reward-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/ask-reddit-reward-testing.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c0b8cf9d3d7e3400"}},{"code_sha256_prefix":"8e058265eaeef576","entry":"get_rewards","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/reward-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/reward-testing.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8e058265eaeef576"}},{"code_sha256_prefix":"1c3dd0abf6725ebd","entry":"load_data","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/md3game.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/md3game.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1c3dd0abf6725ebd"}},{"code_sha256_prefix":"53bbc4b31f140ab9","entry":"load_model","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/ask-reddit-perplexity-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/ask-reddit-perplexity-testing.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"53bbc4b31f140ab9"}},{"code_sha256_prefix":"09fad9c068cf3e99","entry":"load_model","repo":"salt-nlp/unintended-impacts-of-alignment","repo_kind":"official","path":"code/ask-reddit-reward-testing.py","file_url":"https://github.com/salt-nlp/unintended-impacts-of-alignment/blob/HEAD/code/ask-reddit-reward-testing.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"09fad9c068cf3e99"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}