arxiv arXiv cs.AI · 1h ago · src: 2d ago · research

Aethis Eligibility Module uses neuro-symbolic architecture to fix LLM rule evaluation errors

from English

The article documents a failure class in frontier large language models called exception chain collapse, where models incorrectly evaluate nested conditional rules. To address this, the authors present the Aethis Eligibility Module, a neuro-symbolic architecture that combines LLM-authored rules with an SMT-based layer for deterministic execution.

The system ensures consistent rule execution regardless of model drift or prompt format changes.
In a 20-scenario adversarial test on construction insurance, the engine scored 20/20, matching only one of four frontier configurations.
External validation on LegalBench tasks showed the engine was significantly more accurate than three frontier models, with margins up to +41 points against Anthropic models.

The contribution relocates uncertainty from the inference boundary to the specification boundary, making it deliberate and audited for regulated workflows.

Importance 2/3 Beats a top-lab benchmark New harness with differentiators arXiv cs.AI Evaluation & benchmarks Safety & alignment

Read original

media Hugging Face Forums · 3d ago

Jeanbosange releases LIMEN Runtime Audit prototype for inspecting layer-wise activation trajectories

Jeanbosange has published the first public prototype of LIMEN Runtime Audit, an open-source toolkit designed to inspect layer-wise activation trajectories in open-weight language models. The tool treats the sequence of hidden states across layers as a measurable trajectory, providing a reproducible observability layer for comparing internal runtime behavior.

arxiv arXiv cs.CL · 12d ago

UTS bypasses ELOQUENT 2026 Voight-Kampff detectors with structural shifts

Researchers from UTS developed evasion strategies that sweep the top 5 positions on the ELOQUENT 2026 Voight-Kampff leaderboard by exploiting a fundamental asymmetry in detector vulnerability. The study finds that pushing generated text out of the detector's training distribution reliably defeats adversarial detection, whereas mimicking human data fails.

arxiv arXiv cs.CL · 12d ago

Study finds graded entity-familiarity readouts in language models

A study examines activations at the final prompt token in twelve instruction-tuned models from the Bielik, PLLuM, Gemma-4, and Qwen3 families to determine if they can estimate familiarity with an entity before generating an answer. Using a dataset of 1,440 Polish entities, the researchers found that familiarity-probe scores effectively separate real from fabricated entities across all model families.

media Hugging Face Forums · 12d ago · 2 views

Vulcora launches $51.2K challenge to recover hidden trigger in seven SmolLM2-based models

Vulcora has released a public challenge offering a $51,200 bounty for recovering the exact trigger sentence of a hidden backdoor embedded in one of seven open-source language models. The models are derived from HuggingFaceTB/SmolLM2-135M-Instruct and include one teaching model that confesses the mechanism's shape, five decoys, and one with a live undeclared trigger.

arxiv arXiv cs.AI · 13d ago · 1 view

Win by Silence: Deletion Non-Monotonicity and Autonomous Exploitation in LLM Plan Evaluation

This paper investigates a failure mode in staged expected-value scorers for LLM-generated venture routes, where strategic plans can achieve higher scores by becoming less explicit through the deletion of interior transitions.