Attention-based rerankers score documents by aggregating query-to-document attention and subtracting a null-query calibration pass to remove positional and structural bias. Although widely used, this calibration assumes that the null pass is independent of document relevance. We show that modern prompt content, e.g. constraints, instructions, personas, and demonstrations can violate this assumption when it enters the scoring readout, making the null pass relevance-aware rather than null. Through a three-pass decomposition, we find that calibration helps when the null-pass footprint is relevance-agnostic bias, but hurts when it is relevance-tracking signal: short shared constraints survive calibration, whereas long per-query instructions inject signal that subtraction destroys. We propose interpolated null calibration, a training-free modification that controls how much instruction content enters the null baseline. It recovers attention-based reranking on instruction-heavy tasks where standard calibration fails, while preserving calibration’s benefits when the null pass remains relevance-agnostic. On instruction heavy tasks, the recovered rankings surpass generative rerankers such as RankGPT. We also show that in-context demonstrations improve attention-based reranking with little calibration interference, since demonstrations act only through the query pass and leave the null pass unchanged.

Current benchmarks for large language model (LLM) reasoning predominantly focus on mathematical and coding abilities, leaving a gap in evaluating broader reasoning proficiencies. One particular exception is the BIG-Bench dataset, which has served as a crucial benchmark for evaluating the general reasoning capabilities of LLMs, thanks to its diverse set of challenging tasks that allowed for a comprehensive assessment of general reasoning across various skills within a unified framework. However, recent advances in LLMs have led to saturation on BIG-Bench, and its harder version BIG-Bench Hard (BBH). State-of-the-art models achieve near-perfect scores on many tasks in BBH, thus diminishing its utility. To address this limitation, we introduce BIG-Bench Extra Hard (BBEH), a new benchmark designed to push the boundaries of LLM reasoning evaluation. BBEH replaces each task in BBH with a novel task that probes a similar reasoning capability but exhibits significantly increased difficulty. We evaluate various general-purpose and reasoning-specialized models on BBEH and observe an accuracy of 23.9% for the best general-purpose model and 54.2% for the best reasoning-specialized model, indicating substantial room for improvement and highlighting the ongoing challenge of achieving robust general reasoning in LLMs. We release BBEH publicly at: https://github.com/google-deepmind/bbeh.

While self-correction has shown promise in improving LLM outputs in terms of style and quality (e.g. Chen et al., 2023b; Madaan et al.,2023), recent attempts to self-correct logical or reasoning errors often cause correct answers to become incorrect, resulting in worse performances overall (Huang et al., 2023). In this paper, we show that poor self-correction performance stems from LLMs’ inability tofind logical mistakes, rather than their ability to correct a known mistake. Firstly, we benchmark several state-of-the-art LLMs ontheir mistake-finding ability and demonstrate that they generally struggle with the task, even in highly objective, unambiguous cases. Secondly, we test the correction abilities of LLMs – separately from mistake finding – using a backtracking setup that feeds ground truth mistake location information to the model. We show that this boosts downstream task performance across our 5 reasoning tasks, indicating that LLMs’ correction abilities are robust. Finally, we show that it is possible to obtain mistake location information without ground truth labels or in-domain training data. We train a small classifier with out-of-domain data, which exhibits stronger mistake-finding performance than prompting a large model. We release our dataset of LLM-generated logical mistakes, BIG-Bench Mistake, to enable further research into locating LLM reasoning mistakes.

We present Qutr (Query Translator), a smart cross-lingual communication application for the travel domain. Qutr is a real-time messaging app that automatically translates conversations while supporting keyword-to-sentence matching. Qutr relies on querying a database that holds commonly used pre-translated travel-domain phrases and phrase templates in different languages with the use of keywords. The query matching supports paraphrases, incomplete keywords and some input spelling errors. The application addresses common cross-lingual communication issues such as translation accuracy, speed, privacy, and personalization.