Papers › Rethinking the Reranker: Boundary-Aware Evidence Selection for Robust...

Rethinking the Reranker: Boundary-Aware Evidence Selection for Robust Retrieval-Augmented Generation

3 Feb 2026ICMLarXiv:2602.03689added by Syntology

Jiashuo Sun, Pengcheng Jiang, Saizhuo Wang, Jiajun Fan, Heng Wang, Siru Ouyang, Ming Zhong, Yizhu Jiao, Chengsong Huang, Xueqiang Xu, Pengrui Han, Peiran Li, Jiaxin Huang, Ge Liu, Heng Ji, Jiawei Han

Title, abstract, authors and date from arXiv's metadata (CC0); this paper is not in the Papers with Code archive (frozen 2025-07-28).

Retrieval-Augmented Generation (RAG) systems remain brittle under realistic retrieval noise, even when the required evidence appears in the top-K results. A key reason is that retrievers and rerankers optimize solely for relevance, often selecting either trivial, answer-revealing passages or evidence that lacks the critical information required to answer the question, without considering whether the evidence is suitable for the generator. We propose BAR-RAG, which reframes the reranker as a boundary-aware evidence selector that targets the generator's Goldilocks Zone -- evidence that is neither trivially easy nor fundamentally unanswerable for the generator, but is challenging yet sufficient for inference and thus provides the strongest learning signal. BAR-RAG trains the selector with reinforcement learning using generator feedback, and adopts a two-stage pipeline that fine-tunes the generator under the induced evidence distribution to mitigate the distribution mismatch between training and inference. Experiments on knowledge-intensive question answering benchmarks show that BAR-RAG consistently improves end-to-end performance under noisy retrieval, achieving an average gain of 10.3 percent over strong RAG and reranking baselines while substantially improving robustness. Code is publicly avaliable at https://github.com/GasolSun36/BAR-RAG.

PaperPDFCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 26 code samples harvested from 1 repository linked to this paper: 24 ran (of which 4 constructed an object rather than computing a result; 10 with no instrument failure: 6 honoured, 0 violated, 4 with no contract checked; 14 where Syntology's instrument failed) · 2 unverified. Where Syntology's instrument failed: 13 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: found in paper text by Syntology: 26 samples from 1 repository, 24 ran (10 with no instrument failure; 14 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

GasolSun36/BAR-RAG found in paper text by SyntologySyntology: 26 samples, 24 ran (14 instrument-failed) · 2 unverified report

Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

26 samples harvested: 24 ran (of which 4 constructed an object rather than computing a result; 10 with no instrument failure: 6 honoured, 0 violated, 4 with no contract checked; 14 where Syntology's instrument failed) · 2 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

6ran · honoured contract
13ran · our draft was wrong
1ran; the contract check could not run
4ran
2unverified

Licence: 26 of the 26 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from GasolSun36/BAR-RAG. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: 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. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

GasolSun36/BAR-RAG found in paper text by Syntology 26 samples, 24 ran (14 instrument-failed) · 2 unverified
GeneratorRollout filter.py ran · metamorphic tier: deterministic no licence file found · pointer only · f2a668bef88b45e3 · report fixture: returned GeneratorRollout · level recorded 2026-09-22
PreparedSample filter.py ran · metamorphic tier: deterministic no licence file found · pointer only · 433f1cac108e3a31 · report fixture: returned PreparedSample · level recorded 2026-09-22
SampleResult filter.py ran · metamorphic tier: deterministic no licence file found · pointer only · 6a0889d51d7f3b3e · report fixture: returned SampleResult · level recorded 2026-09-22
SelectorRollout filter.py ran · metamorphic tier: deterministic no licence file found · pointer only · 6c2105013e64ad11 · report fixture: returned SelectorRollout · level recorded 2026-09-22
build_generator_prompt_prefix filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 0616ef45c981d85c · report fixture: signature question:str/0/str · returned str · level recorded 2026-09-22
build_generator_prompt_with_budget filter.py ran · our draft was wrong no licence file found · pointer only · 09e3e7200f10e115 · report fixture: signature tokenizer:MockTokenizer/0/?, question:str/0/str, selected_ctxs:list/1/dict, max_prompt_tokens:int/0/int, reserve_for_generation_tokens:int/0/int · returned str · level recorded 2026-09-22
build_selector_conversation filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · a010a2ea8a5ecf9f · report fixture: signature question:str/0/str, retrieved_content:str/0/str · returned list · level recorded 2026-09-22
cite_reward_peaked filter.py ran · honoured contract no licence file found · pointer only · fde56a8a1f8a1fac · report fixture: signature predict_str:str/0/str, target_cite_count:int/0/int, max_doc_id:int/0/int · returned float · level recorded 2026-09-22
compute_generator_reward filter.py ran · our draft was wrong no licence file found · pointer only · d1600919970d1d70 · report fixture: signature predict_str:str/0/str, golden_answers:list/1/str, max_doc_id:int/0/int, target_cite_count:int/0/int, acc_weight:float/0/float · returned dict · level recorded 2026-09-22
compute_hbs_score filter.py ran · honoured contract fingerprinted no licence file found · pointer only · cfe808026873ade9 · report fixture: signature mean:float/0/float, var:float/0/float, sigma:float/0/float, center:float/0/float · returned float · level recorded 2026-09-22
count_tokens in 39 papers filter.py ran · honoured contract no licence file found · pointer only · 0335852d729ef632 · report fixture: signature tokenizer:MockTokenizer/0/?, text:str/0/str · returned int · level recorded 2026-09-22
count_unique_doc_cites_in_think filter.py ran · honoured contract fingerprinted no licence file found · pointer only · bdd51fcd3a78301f · report fixture: signature predict_str:str/0/str, max_doc_id:int/0/int · returned int · level recorded 2026-09-22
extract_answer_tag filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · ac37ef9234f428e1 · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-22
extract_selector_answer filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 102728db38c7b95b · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-22
extract_selector_think filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 5dc60c2aea0bb519 · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-22
extract_think_tag filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 5457a476740b4a4f · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-22
format_doc_block filter.py ran · our draft was wrong no licence file found · pointer only · 912b9a1b8d856a37 · report fixture: signature i:int/0/int, title:str/0/str, content:str/0/str · returned str · level recorded 2026-09-22
format_top_n_docs filter.py ran · our draft was wrong no licence file found · pointer only · ab9ac8a021f0a8bf · report fixture: signature ctxs:list/1/dict, n:int/0/int · returned str · level recorded 2026-09-22
normalize_answer in 304 papers filter.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 85fc7d4bd5e8711a · report fixture: signature s:str/0/str · returned str · level recorded 2026-09-22
parse_selector_output_to_doc_indices filter.py ran · honoured contract no licence file found · pointer only · c73169833a02a1de · report fixture: signature predict_str:str/0/str, num_candidates:int/0/int, default_k:int/0/int · returned list · level recorded 2026-09-22
pass_hbs_filter filter.py ran; the contract check could not run (RAISES) · metamorphic tier: deterministic no licence file found · pointer only · 7ee1876e2e7de116 · report fixture: signature acc_list:list/1/float, mean:float/0/float, var:float/0/float, mean_min:float/0/float, mean_max:float/0/float, var_min:float/0/float, require_cross_50:bool/0/bool · returned bool · level recorded 2026-09-22
prepare_samples filter.py ran · our draft was wrong no licence file found · pointer only · 7bc562289b188468 · report fixture: signature samples:list/1/dict · returned list · level recorded 2026-09-22
squad_f1_score filter.py ran · honoured contract fingerprinted no licence file found · pointer only · f52624c81dd65473 · report fixture: signature prediction:str/0/str, ground_truth:str/0/str · returned float · level recorded 2026-09-22
truncate_to_tokens filter.py ran · our draft was wrong no licence file found · pointer only · e5bd9293acb23884 · report fixture: signature tokenizer:MockTokenizer/0/?, text:str/0/str, max_tokens:int/0/int · returned str · level recorded 2026-09-22
VLLMClient filter.py unverified no licence file found · pointer only · a6c1044cbe0d3f31 · report
process_batch_samples filter.py unverified no licence file found · pointer only · 997b974a1cb1d2db · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for BAR-RAG (graph method: BAR-RAG). 26 bricks; individually on a fixture: 24 ran (of which 4 constructed an object rather than computing a result; 10 with no instrument failure: 6 honoured, 0 violated, 4 with no contract checked; 14 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

  1. process_batch_samples entry · level 0 · unverified · 997b974a1cb1d2db
  2. GeneratorRollout helper · level 1 · ran · f2a668bef88b45e3
  3. SampleResult helper · level 1 · ran · 6a0889d51d7f3b3e
  4. SelectorRollout helper · level 1 · ran · 6c2105013e64ad11
  5. VLLMClient helper · level 0 · unverified · a6c1044cbe0d3f31
  6. build_generator_prompt_with_budget data_transform · level 1 · ran · our draft was wrong · 09e3e7200f10e115
  7. compute_generator_reward metric · level 1 · ran · our draft was wrong · d1600919970d1d70
  8. compute_hbs_score metric · level 1 · ran · honoured contract · cfe808026873ade9
  9. extract_selector_answer helper · level 1 · ran · our draft was wrong · 102728db38c7b95b
  10. extract_selector_think helper · level 1 · ran · our draft was wrong · 5dc60c2aea0bb519
  11. parse_selector_output_to_doc_indices data_transform · level 1 · ran · honoured contract · c73169833a02a1de
  12. pass_hbs_filter metric · level 1 · ran; the contract check could not run (RAISES) · 7ee1876e2e7de116
  13. prepare_samples data_transform · level 1 · ran · our draft was wrong · 7bc562289b188468
  14. build_generator_prompt_prefix helper · level 1 · ran · our draft was wrong · 0616ef45c981d85c
  15. count_tokens helper · level 1 · ran · honoured contract · 0335852d729ef632
  16. format_doc_block helper · level 1 · ran · our draft was wrong · 912b9a1b8d856a37
  17. truncate_to_tokens helper · level 1 · ran · our draft was wrong · e5bd9293acb23884
  18. cite_reward_peaked metric · level 1 · ran · honoured contract · fde56a8a1f8a1fac
  19. count_unique_doc_cites_in_think helper · level 1 · ran · honoured contract · bdd51fcd3a78301f
  20. extract_answer_tag helper · level 1 · ran · our draft was wrong · ac37ef9234f428e1
  21. normalize_answer helper · level 1 · ran · our draft was wrong · 85fc7d4bd5e8711a
  22. squad_f1_score metric · level 1 · ran · honoured contract · f52624c81dd65473
  23. PreparedSample helper · level 1 · ran · 433f1cac108e3a31
  24. build_selector_conversation data_transform · level 1 · ran · our draft was wrong · a010a2ea8a5ecf9f
  25. format_top_n_docs data_transform · level 1 · ran · our draft was wrong · ab9ac8a021f0a8bf
  26. extract_think_tag helper · level 1 · ran · our draft was wrong · 5457a476740b4a4f

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2602.03689") returns these assemblies with each brick's role and wiring (how to connect).

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2602.03689, was issued in February 2026, after that date, so the archive has no leaderboard rows for it.

Placed on leaderboards by Syntology Syntology

Syntology's extractor, the model Claude Sonnet 4.5, judged this paper's tables to report on 2 leaderboards (a model's judgement, not a result) and has not placed the paper on any of them: Open-Domain Question Answering · TriviaQA (refused by a rule: one configuration gave two values for one column); Question Answering · TriviaQA (refused by a rule: one configuration gave two values for one column). What is not shown.

Syntology has checked 8,886 of the 18,637 papers on this site that are newer than the archive (for 3,717 of them no archive leaderboard matched the paper's tables, so there was nothing further to check); 733 were read and have nothing to place (arXiv has no HTML version of the paper, or that version has no tables), 41 could not be read (the extractor's reply could not be parsed), and results from the other 8,977 appear after they are checked.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections