Picture for Melanie Sclar

Melanie Sclar

The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning

Add code
Dec 04, 2023
Figure 1 for The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning
Figure 2 for The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning
Figure 3 for The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning
Figure 4 for The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning
Viaarxiv icon

FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions

Add code
Oct 31, 2023
Figure 1 for FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions
Figure 2 for FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions
Figure 3 for FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions
Figure 4 for FANToM: A Benchmark for Stress-testing Machine Theory of Mind in Interactions
Viaarxiv icon

Quantifying Language Models' Sensitivity to Spurious Features in Prompt Design or: How I learned to start worrying about prompt formatting

Add code
Oct 17, 2023
Viaarxiv icon

Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement

Add code
Oct 12, 2023
Figure 1 for Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement
Figure 2 for Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement
Figure 3 for Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement
Figure 4 for Phenomenal Yet Puzzling: Testing Inductive Reasoning Capabilities of Language Models with Hypothesis Refinement
Viaarxiv icon

Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker

Add code
Jun 01, 2023
Figure 1 for Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker
Figure 2 for Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker
Figure 3 for Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker
Figure 4 for Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker
Viaarxiv icon

Faith and Fate: Limits of Transformers on Compositionality

Jun 01, 2023
Figure 1 for Faith and Fate: Limits of Transformers on Compositionality
Figure 2 for Faith and Fate: Limits of Transformers on Compositionality
Figure 3 for Faith and Fate: Limits of Transformers on Compositionality
Figure 4 for Faith and Fate: Limits of Transformers on Compositionality
Viaarxiv icon

Referee: Reference-Free Sentence Summarization with Sharper Controllability through Symbolic Knowledge Distillation

Add code
Oct 25, 2022
Figure 1 for Referee: Reference-Free Sentence Summarization with Sharper Controllability through Symbolic Knowledge Distillation
Figure 2 for Referee: Reference-Free Sentence Summarization with Sharper Controllability through Symbolic Knowledge Distillation
Figure 3 for Referee: Reference-Free Sentence Summarization with Sharper Controllability through Symbolic Knowledge Distillation
Figure 4 for Referee: Reference-Free Sentence Summarization with Sharper Controllability through Symbolic Knowledge Distillation
Viaarxiv icon