Papers › When RLHF Fails: A Mechanistic Taxonomy of Reward Hacking, Collapse, and Evaluator Gaming

When RLHF Fails: A Mechanistic Taxonomy of Reward Hacking, Collapse, and Evaluator Gaming

2 Jun 2026arXiv:2606.03238added by Syntology

Zelalem Abahana, David Evans, Satish Mahadevan Srinivasan, Matjaz Gams

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

RLHF evaluation should track how failures emerge, where they localize, and which warning signals appear before external quality degrades. We study this problem with a compact RLHF pipeline built for this paper, including PPO, DPO, uncertainty-penalized PPO (UP-PPO), reward-model uncertainty, approximate policy drift, diversity and repetition diagnostics, and two external LLM judges. Rather than treating reward hacking as a single terminal event, we classify matched checkpoint and prompt-level transitions by the directions of learned reward R_phi, judge scores R_dag and R2_dag, and their average R_dag. The main empirical findings are that aggressive PPO produces the clearest localized reward-hacking signal, UP-PPO reduces but does not eliminate that signal, row-level diagnostics reveal failures hidden by checkpoint averages, and pre-transition features partially anticipate future localized reward hacking. The central conclusion is methodological: RLHF failures are training dynamics that can be classified, localized, and partially anticipated, not only final-model pathologies. The repository is available at github.com/zabahana/rlhf-failure-modes-diagnostics. The pipeline is also deployed as a live interactive web demo for model comparison and diagnostic views at rlhf-failures.zelalem.ai.

PaperPDFCode Syntology ran

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

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2606.03238")

Code

Syntology Ran 8 of 11 code samples harvested from 1 repository linked to this paper; 3 have no recorded run. Of those that ran: 8 ran with no contract checked.

By repository: found in paper text by Syntology: 11 samples from 1 repository, 8 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

zabahana/rlhf-failure-modes-diagnostics found in paper text by Syntology 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

11 samples harvested; 8 ran; 0 honoured the contract we drafted; 3 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

8ran
3unverified

Licence: 11 of the 11 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 zabahana/rlhf-failure-modes-diagnostics. “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. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (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 from a repository marked official are listed first. 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 four 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; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). 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.

align_preference_for_dpo zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/dpo_data.py found in paper text by Syntology ran licence not identified · pointer only · cdded901c0901a52 · report
align_row_for_dpo zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/dpo_data.py found in paper text by Syntology ran licence not identified · pointer only · 8fe1ea343a43d68f · report
build_examples zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/data_hh.py found in paper text by Syntology ran licence not identified · pointer only · 3553f836c857307e · report
classify_proxy_judge_transition zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/failure_modes.py found in paper text by Syntology ran licence not identified · pointer only · 1618500fd1e2b241 · report
classify_transition zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/failure_modes.py found in paper text by Syntology ran licence not identified · pointer only · ef51981fecd4fe12 · report
detect_evaluator_gaming zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/failure_modes.py found in paper text by Syntology ran licence not identified · pointer only · 26759ed9d194a198 · report
mock_judge_score zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/judge_mock.py found in paper text by Syntology ran fingerprinted licence not identified · pointer only · 6132151deece66df · report
split_hh_row zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/data_hh.py found in paper text by Syntology ran licence not identified · pointer only · 9f69b89fd7919b4c · report
anthropic_helpfulness_1_to_10 zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/judge_anthropic.py found in paper text by Syntology unverified licence not identified · pointer only · 0026e1d9eb2694f3 · report
openai_helpfulness_1_to_10 zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/judge_openai.py found in paper text by Syntology unverified licence not identified · pointer only · 0073e7e0932efec7 · report
resolve_second_judge zabahana/rlhf-failure-modes-diagnostics/src/rlhf_pipeline/judge_resolution.py found in paper text by Syntology unverified licence not identified · pointer only · dc51164f23e8b0da · report

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2606.03238, was issued in June 2026, after that date, so the archive has no leaderboard rows for it.

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