DRAGged into Conflicts: Detecting and Addressing Conflicting Sources in Search-Augmented LLMs
Paper
•
2506.08500
•
Published
•
7
Roee Aharoni
roeeaharoni
AI & ML interests
Natural Language Processing
Recent Activity
upvoted
a
paper
4 months ago
RefVNLI: Towards Scalable Evaluation of Subject-driven Text-to-image
Generation
upvoted
a
paper
5 months ago
Inside-Out: Hidden Factual Knowledge in LLMs
Organizations
upvoted
a
paper
2 months ago
upvoted
a
paper
4 months ago
upvoted
a
paper
5 months ago
liked
a
dataset
about 1 year ago
reacted to
gsarti's
post with 🤗
over 1 year ago
Post
🔍 Today's pick in Interpretability & Analysis of LMs: A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains by
@alonjacovi
@yonatanbitton
B. Bohnet J. Herzig
@orhonovic
M. Tseng M. Collins
@roeeaharoni
@mega
This work introduces a new methodology for human verification of reasoning chains and adopts it to annotate a dataset of chain-of-thought reasoning chains produced by 3 LMs. The annotated dataset, REVEAL, can be used to benchmark automatic verifiers of reasoning in LMs.
In their analysis, the authors find that LM-produced CoTs generally contain faulty steps, often leading to incorrect automatic verification. In particular, CoT-generating LMs are found to produce non-attributable reasoning steps often, and reasoning verifiers generally struggle to verify logical correctness.
📄 Paper: A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains (2402.00559)
🔡 Dataset: google/reveal
This work introduces a new methodology for human verification of reasoning chains and adopts it to annotate a dataset of chain-of-thought reasoning chains produced by 3 LMs. The annotated dataset, REVEAL, can be used to benchmark automatic verifiers of reasoning in LMs.
In their analysis, the authors find that LM-produced CoTs generally contain faulty steps, often leading to incorrect automatic verification. In particular, CoT-generating LMs are found to produce non-attributable reasoning steps often, and reasoning verifiers generally struggle to verify logical correctness.
📄 Paper: A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains (2402.00559)
🔡 Dataset: google/reveal
authored
a
paper
over 1 year ago
liked
a
model
almost 2 years ago
authored
a
paper
about 2 years ago
updated
a
model
over 2 years ago
liked
a
model
over 2 years ago
updated
a
model
almost 3 years ago