Claims Browser

RATIO is constructed from millions of full-text scientific papers across CS literature using discourse-marker distant supervision extended to corpus-scale retrieval, combined with LLM and human vetting

"RATIO is constructed from millions of full-text scientific papers across CS literature via a general recipe that extends discourse-marker distant supervision - previously used only for classification - to corpus-scale retrieval, combined with extensive LLM and human vetting."

Operation-specific fine-tuning substantially boosts retriever performance but leaves much room for further improvements

"Experiments show that operation-specific fine-tuning substantially boosts retrievers but leaves much room for further improvements."

LLMs show increasingly poor performance as input size approaches realistic corporate communication scales of 230,000+ documents

"We evaluate five LLMs on CB, revealing increasingly poor performance as input size approaches realistic scales."

LLMs are increasingly capable of answering complex questions about enterprise-scale document collections

"LLMs are increasingly able to answer complex questions about enterprise-scale document collections."

Existing AI systems have unclear inclusivity for blind and deafblind users accessing functionality through Braille

"it is unclear whether existing AI systems are inclusive enough for blind and deafblind users to access the same functionality through Braille"

There is a persistent gap between LLM capabilities in print-English versus Braille accessibility

"The results reveal a persistent gap between print-English capability and Braille accessibility"

Braille understanding and expression are asymmetric in LLMs, with Grade 2 being especially fragile on the input side compared to Grade 1

"Braille understanding and expression are asymmetric, where Grade 2 is especially fragile on the input side compared to Grade 1"

Fully Braille requests further reduce LLM performance beyond the existing accessibility gap

"fully Braille requests further reduce performance"

Double-difference designs used to audit LLM judge bias are not properly identified on bounded rating scales, confounding differential preference with differential attenuation

"We show that this endpoint is not identified on the scale that reports it. Each term of the double difference is censored by its own share, so the observed statistic confounds differential preference with differential attenuation"

A severity shift common to both responses can manufacture a spurious interaction effect whenever two responses censor it unequally due to unequal distances from scale bounds

"a severity shift common to both responses manufactures an interaction whenever the two censor it unequally, as unequal distances from the bounds make them, exactly where good stimuli place them"

In a pre-registered audit of a pedagogy judge with 990 calls, the effect of stated learner profile on scaffolding preference was null at +0.085 points

"The registered primary endpoint, the effect of a stated learner profile on the judge's scaffolding preference, is null: $+0.085$ points (95\% BCa $[-0.167, +0.353]$, $p = 0.684$)"

A nominally significant interaction effect of +0.378 in the audit is not actually identified as preference, with 79-85% of it reproducible from severity shift and scale floor alone without any differential preference

"The audit's one nominally significant interaction, $+0.378$ ($p = 0.002$), is not identified as preference: a construction containing zero differential preference reproduces 79 to 85\% of it from the observed severity shift and the scale floor alone"

The contribution of the censoring mechanism to observed interaction effects is measurable from an audit's own ratings and can be derived in closed form

"We derive the mechanism in closed form and show that its contribution is measurable from an audit's own ratings"

High similarity between first-visit and return frames in video world models does not necessarily indicate scene memory, as the intervening rollout may have changed very little

"High similarity between first-visit and return frames does not necessarily show that a video world model remembered the scene; the intervening rollout may simply have changed very little."

R2M-Bench's Overall NMR metric correlates with human consistency judgments at Spearman's ρ=0.547

"Overall NMR correlates with human consistency judgments at Spearman's $ρ=0.547$ (95\% CI $[0.45,0.63]$)."

Relative calibration substantially reduces the slow-motion shortcut in video world model evaluation, with within-model correlation magnitude of 0.072 compared to 0.207 for raw revisit similarity

"Its within-model correlation magnitude with generated motion is $0.072$, compared with $0.207$ for raw revisit similarity, indicating that relative calibration substantially reduces the slow-motion shortcut."

DreamX-World-Memo achieves the highest Overall NMR among evaluated video models

"DreamX-World-Memo achieves the highest Overall NMR among the evaluated video models."

Industrial sites contain large volumes of read-only telemetry, but few benchmarks specify how to compile these records into executable multi-turn agent tasks

"Industrial sites contain large volumes of read-only telemetry, but few benchmarks specify how to compile these records into executable multi-turn agent tasks."

BTS-AgentBench provides a telemetry-to-episode construction method that successfully produces 532 rows with multiple advanced features including clarification, goal revision, and evidence attribution

"The 532-row release adds clarification, goal revision, timestamp policy, quality-gated reporting, and evidence attribution while preserving the source computation and split."