Picture for Ansh Radhakrishnan

Ansh Radhakrishnan

Debating with More Persuasive LLMs Leads to More Truthful Answers

Add code
Feb 15, 2024
Figure 1 for Debating with More Persuasive LLMs Leads to More Truthful Answers
Figure 2 for Debating with More Persuasive LLMs Leads to More Truthful Answers
Figure 3 for Debating with More Persuasive LLMs Leads to More Truthful Answers
Figure 4 for Debating with More Persuasive LLMs Leads to More Truthful Answers
Viaarxiv icon

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Add code
Jan 17, 2024
Viaarxiv icon

Question Decomposition Improves the Faithfulness of Model-Generated Reasoning

Add code
Jul 25, 2023
Figure 1 for Question Decomposition Improves the Faithfulness of Model-Generated Reasoning
Figure 2 for Question Decomposition Improves the Faithfulness of Model-Generated Reasoning
Figure 3 for Question Decomposition Improves the Faithfulness of Model-Generated Reasoning
Figure 4 for Question Decomposition Improves the Faithfulness of Model-Generated Reasoning
Viaarxiv icon

Measuring Faithfulness in Chain-of-Thought Reasoning

Jul 17, 2023
Figure 1 for Measuring Faithfulness in Chain-of-Thought Reasoning
Figure 2 for Measuring Faithfulness in Chain-of-Thought Reasoning
Figure 3 for Measuring Faithfulness in Chain-of-Thought Reasoning
Figure 4 for Measuring Faithfulness in Chain-of-Thought Reasoning
Viaarxiv icon