Show, Don't Tell: Evaluating Large Language Models Beyond Textual Understanding with ChildPlay

• URL: http://arxiv.org/abs/2407.11068v3
• Date: Sun, 18 Aug 2024 19:44:35 GMT
• Title: Show, Don't Tell: Evaluating Large Language Models Beyond Textual Understanding with ChildPlay
• Authors: Gonçalo Hora de Carvalho, Oscar Knap, Robert Pollice,
• Abstract summary: We use games like Tic-Tac-Toe, Connect Four, and Battleship to assess strategic thinking and decision-making. Despite their proficiency on standard benchmarks, GPT-3.5 and GPT-4's abilities to play and reason about fully observable games without pre-training is mediocre.
• Score: 0.0
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We explore the hypothesis that LLMs, such as GPT-3.5 and GPT-4, possess broader cognitive functions, particularly in non-linguistic domains. Our approach extends beyond standard linguistic benchmarks by incorporating games like Tic-Tac-Toe, Connect Four, and Battleship, encoded via ASCII, to assess strategic thinking and decision-making. To evaluate the models' ability to generalize beyond their training data, we introduce two additional games. The first game, LEGO Connect Language (LCL), tests the models' capacity to understand spatial logic and follow assembly instructions. The second game, the game of shapes, challenges the models to identify shapes represented by 1s within a matrix of zeros, further testing their spatial reasoning skills. This "show, don't tell" strategy uses games instead of simply querying the models. Our results show that despite their proficiency on standard benchmarks, GPT-3.5 and GPT-4's abilities to play and reason about fully observable games without pre-training is mediocre. Both models fail to anticipate losing moves in Tic-Tac-Toe and Connect Four, and they are unable to play Battleship correctly. While GPT-4 shows some success in the game of shapes, both models fail at the assembly tasks presented in the LCL game. These results suggest that while GPT models can emulate conversational proficiency and basic rule comprehension, their performance in strategic gameplay and spatial reasoning tasks is very limited. Importantly, this reveals a blind spot in current LLM benchmarks that we highlight with our gameplay benchmark suite ChildPlay (https://github.com/child-play-neurips/child-play). Our findings provide a cautionary tale about claims of emergent intelligence and reasoning capabilities of LLMs that are roughly the size of GPT-3.5 and GPT-4.

Related papers

• Causal World Representation in the GPT Model [4.629721760278161]
  generative pre-trained transformer (GPT) models are tested on real-world games played with the intention of winning. We find that GPT models tend to generate next moves that adhere to the game rules for sequences for which the attention mechanism encodes a causal structure with high confidence. In general, in cases for which the GPT model generates moves that do not adhere to the game rules, it also fails to capture any causal structure.
  arXiv  Detail & Related papers  (2024-12-10T12:05:03Z)
• Are Large Language Models Strategic Decision Makers? A Study of Performance and Bias in Two-Player Non-Zero-Sum Games [56.70628673595041]
  Large Language Models (LLMs) have been increasingly used in real-world settings, yet their strategic decision-making abilities remain largely unexplored. This work investigates the performance and merits of LLMs in canonical game-theoretic two-player non-zero-sum games, Stag Hunt and Prisoner Dilemma. Our structured evaluation of GPT-3.5, GPT-4-Turbo, GPT-4o, and Llama-3-8B shows that these models, when making decisions in these games, are affected by at least one of the following systematic biases.
  arXiv  Detail & Related papers  (2024-07-05T12:30:02Z)
• Adaptable Logical Control for Large Language Models [68.27725600175013]
  Ctrl-G is an adaptable framework that facilitates tractable and flexible control of model generation at inference time. We show that Ctrl-G, when applied to a TULU2-7B model, outperforms GPT3.5 and GPT4 on the task of interactive text editing.
  arXiv  Detail & Related papers  (2024-06-19T23:47:59Z)
• GameBench: Evaluating Strategic Reasoning Abilities of LLM Agents [4.209869303518743]
  We introduce GameBench, a cross-domain benchmark for evaluating strategic reasoning abilities of large language models. Our evaluations use GPT-3 and GPT-4 in their base form along with two scaffolding frameworks designed to enhance strategic reasoning ability: Chain-of-Thought (CoT) prompting and Reasoning Via Planning (RAP) Our results show that none of the tested models match human performance, and at worst GPT-4 performs worse than random action.
  arXiv  Detail & Related papers  (2024-06-07T00:28:43Z)
• DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models [92.6951708781736]
  This work proposes a comprehensive trustworthiness evaluation for large language models with a focus on GPT-4 and GPT-3.5. We find that GPT models can be easily misled to generate toxic and biased outputs and leak private information. Our work illustrates a comprehensive trustworthiness evaluation of GPT models and sheds light on the trustworthiness gaps.
  arXiv  Detail & Related papers  (2023-06-20T17:24:23Z)
• Gpt-4: A Review on Advancements and Opportunities in Natural Language Processing [0.0]
  Generative Pre-trained Transformer 4 (GPT-4) is the fourth-generation language model in the GPT series, developed by OpenAI. GPT-4 has a larger model size (more than one trillion), better multilingual capabilities, improved contextual understanding, and reasoning capabilities than GPT-3. Some of the potential applications of GPT-4 include chatbots, personal assistants, language translation, text summarization, and question-answering.
  arXiv  Detail & Related papers  (2023-05-04T22:46:43Z)
• Analyzing the Performance of GPT-3.5 and GPT-4 in Grammatical Error Correction [28.58384091374763]
  GPT-3 and GPT-4 models are powerful, achieving high performance on a variety of Natural Language Processing tasks. We perform experiments testing the capabilities of a GPT-3.5 model (text-davinci-003) and a GPT-4 model (gpt-4-0314) on major GEC benchmarks. We report the performance of our best prompt on the BEA-2019 and JFLEG datasets, finding that the GPT models can perform well in a sentence-level revision setting.
  arXiv  Detail & Related papers  (2023-03-25T03:08:49Z)
• Sparks of Artificial General Intelligence: Early experiments with GPT-4 [66.1188263570629]
  GPT-4, developed by OpenAI, was trained using an unprecedented scale of compute and data. We demonstrate that GPT-4 can solve novel and difficult tasks that span mathematics, coding, vision, medicine, law, psychology and more. We believe GPT-4 could reasonably be viewed as an early (yet still incomplete) version of an artificial general intelligence (AGI) system.
  arXiv  Detail & Related papers  (2023-03-22T16:51:28Z)
• A Comprehensive Capability Analysis of GPT-3 and GPT-3.5 Series Models [71.42197262495056]
  GPT series models have gained considerable attention due to their exceptional natural language processing capabilities. We select six representative models, comprising two GPT-3 series models and four GPT-3.5 series models. We evaluate their performance on nine natural language understanding (NLU) tasks using 21 datasets. Our experiments reveal that the overall ability of GPT series models on NLU tasks does not increase gradually as the models evolve.
  arXiv  Detail & Related papers  (2023-03-18T14:02:04Z)
• Prompting GPT-3 To Be Reliable [117.23966502293796]
  This work decomposes reliability into four facets: generalizability, fairness, calibration, and factuality. We find that GPT-3 outperforms smaller-scale supervised models by large margins on all these facets.
  arXiv  Detail & Related papers  (2022-10-17T14:52:39Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.