AI Papers of the Week
Every paper worth reading in AI, hand-picked one week at a time.
Humpback (Self-Alignment with Instruction Backtranslation)
Meta's Humpback automatically generates instruction-tuning data by back-translating web text into plausible instructions.
Shepherd
Meta's Shepherd is a 7B language model specifically tuned to critique model outputs and suggest refinements.
Political Biases in NLP Models
Develops methods to measure political and media biases in LLMs and their downstream effects.
Studying LLM Generalization with Influence Functions
Anthropic scales influence functions to LLMs up to 52B parameters to investigate generalization patterns.
Synthetic Data Reduces Sycophancy
Google shows that fine-tuning on simple synthetic data can significantly reduce LLM sycophancy.
Trustworthy LLMs
Presents a comprehensive framework of categories for assessing LLM trustworthiness.
Open Problems and Limitations of RLHF
A comprehensive survey of open problems and fundamental limitations of RLHF as an alignment approach.
The Hydra Effect
DeepMind shows that language models exhibit self-repairing behavior when attention heads are ablated.
Self-Check
Explores LLM capacity for self-checking on complex reasoning tasks requiring multi-step and non-linear thinking.
Universal Adversarial LLM Attacks
Finds universal and transferable adversarial attacks that cause aligned models like ChatGPT and Bard to generate objectionable behaviors.
Survey of Aligned LLMs
A comprehensive overview of alignment approaches covering data, training, and evaluation.
Llama 2
Meta's open-weight foundation model family with chat-tuned variants ranging from 7B to 70B parameters.
How is ChatGPT's Behavior Changing Over Time?
Evaluates GPT-3.5 and GPT-4 over months to show significant behavioral drift in deployed systems.
Measuring Faithfulness in Chain-of-Thought Reasoning
Anthropic's investigation into whether CoT reasoning actually reflects the model's internal decision process.
Challenges & Application of LLMs
A comprehensive enumeration of open challenges and application domains for LLMs.
FLASK
Proposes fine-grained evaluation of LLMs decomposed into 12 alignment skill sets.
Claude 2
Anthropic's second-generation LLM with a detailed model card on safety, alignment, and capabilities.
Robots That Ask for Help
A framework for calibrating LLM-based robot planners so they ask for help when uncertain.
An Overview of Catastrophic AI Risks
Dan Hendrycks' comprehensive overview of catastrophic AI risk categories.
LMFlow
An extensible and lightweight toolkit for fine-tuning and inference of large foundation models.
Reliability of Watermarks for LLMs
Studies whether watermarks survive human rewriting and LLM paraphrasing.
Concept Scrubbing in LLM (LEACE)
Least-squares Concept Erasure - erases a target concept from every layer of a neural network.
Direct Preference Optimization (DPO)
Rafailov et al.'s simpler alternative to RLHF that rivals full RL-based alignment.
LIMA
Meta's 65B LLaMA fine-tuned on just 1,000 curated examples - showing alignment needs less data than believed.