Tag
This paper proposes measuring concept content in text using LLM internal activations via linear probes and RFM concept vectors, applied to ESG classification. The best linear probe approaches fine-tuned classifier accuracy without task-specific fine-tuning and outperforms the model's own output, showing activations carry concept content beyond responses.
Scope3Trace proposes an evidence-grounded information extraction framework that leverages large language models to identify and extract Scope 3 greenhouse gas emissions from sustainability reports, contributing a dual-level multimodal dataset and achieving high extraction accuracy.
Due to malicious layoffs and stock option cancellation disputes under its VIE structure, Xiaohongshu may see its Hong Kong IPO process affected, resulting in significant discounts in valuation and delays in listing timeline.
This paper proposes a deterministic climate-risk intelligence framework integrating orchestration, anomaly detection, and imbalance-aware ensemble learning for auditable ESG validation, addressing fragmented Scope 1-3 reporting data.
A Danish pension fund has blacklisted SpaceX, citing concerns over 'catastrophic governance' related to Elon Musk's leadership and controversial behavior.
MIT researchers propose ESGLens, a RAG framework that extracts structured ESG data from PDF reports and predicts environmental scores with 0.48 Pearson correlation against LSEG benchmarks.