Unverified50% confidenceFactExact time
On the three benchmarks UNSW-NB15, CIC-IDS2017, and EMBER2024, the paper compares against calibrated XGBoost, Random-Forest, MLP, and Kitsune, and measures a two-stage cascade of "Suricata (52,674 Emerging-Threats rules) → VX-One" running over five days on CIC-IDS2017.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
Unverifiedd1-3B on text tasks: SQuAD 2.0 85.3, Civil Comments 93.0, MASSIVE intent 87.3, PubMedQA 66.0, BoolQ 86.7, XNLI 85.0, PAWS-X 76.9, averaging 82.9 across the seven.60% similarUnverifiedThe authors obtained 2,213 generation trajectories, 2,108 planning trajectories, and 2,013 rounds of validation improvement data from 4,003 candidates, which were used for supervised fine-tuning of Qwen3-8B-27B.58% similarUnverifiedThe research team conducted zero-shot evaluations of off-the-shelf models on 15 long-context tasks, with test subjects including Gemma-4-31B and Qwen-3.6-27B.58% similarUnverifiedIn testing on the full BEIR SciFact dataset (5,183 documents, 300 queries), Qwen3-Embedding-8B ranked first with 0.7953 nDCG@10, with EmbeddingGemma-300M close behind at 0.7861.58% similarUnverifiedThe scoring script checks three hard metrics: file count (default threshold: 8 files), lines changed (default threshold: 200 lines), and whether any blocked paths appear in the diff.58% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084097API
curl https://kongchang.com/api/v1/knowledge/claims/1084097MCP
get_claim(id=1084097)