Unverified50% confidenceFactExact time
Probing layer 2 (mean + standard deviation pooling) of EmbeddingGemma-2 raises VoxPopuli gender clustering from 0.05% at the final output to 66.6% (layer 3)/above 65% (layer 2), and IEMOCAP gender classification from 66.4% to 83.0%.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCounterfactual mDPO improved METEOR from 58.10% to 85.30%, reduced Self-BLEU-2 from 58.79% to 52.88%, and improved Distinct-3 from 6.68% to 7.81%.59% similarUnverifiedAfter incorporating EvoMem, all baseline agents showed broad improvement, with Chain Level gains notably larger than Step Level gains: in the Terminal scenario, Chain Level improved by approximately 6 percentage points (Step Level only 2.4%), in the SWE scenario by 2 percentage points, and in the Persona-Mem scenario by 3.2 percentage points.59% similarUnverifiedLogic42's laboratory simulation of a standard enterprise fine-tuning pipeline (SFT+LoRA, five rounds of recursive looping, 100,000 synthetic technical support tickets) shows that the human ground-truth ratio dropped from 100% at Gen 0 to 0% at Gen 4, domain precision dropped from 94.6% to 44.3%, rare-tail recall dropped from 88.2% to 11.2%, and the hallucination rate rose from 3.2% to 52.4%.58% similarUnverified遗传因素约占运动表现变异的50-80%（基于双胞胎研究）57% similarUnverifiedThe model with the extended vocabulary improved its pass rate on HumanEval from 45.7% to 52.4%.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1108278API
curl https://kongchang.com/api/v1/knowledge/claims/1108278MCP
get_claim(id=1108278)