Framing Effects in Independent-Agent Large Language Models: A Cross-Family Behavioral Analysis

Zice Wang; Zhenyu Zhang

cs.CLMar 2, 2026

Framing Effects in Independent-Agent Large Language Models: A Cross-Family Behavioral Analysis

Authors: Zice Wang, Zhenyu Zhang

Organizations: Northeastern University, Shenyang 110819, Liaoning, China. · Beijing Institute of Technology, Beijing, 100081, China.

Abstract

In many real-world applications, large language models (LLMs) operate as independent agents without interaction, thereby limiting coordination. In this setting, we examine how prompt framing influences decisions in a threshold voting task involving individual-group interest conflict. Two logically equivalent prompts with different framings were tested across diverse LLM families under isolated trials. Results show that prompt framing significantly influences choice distributions, often shifting preferences toward risk-averse options. Surface linguistic cues can even override logically equivalent formulations. This suggests that observed behavior reflects a tendency consistent with a preference for instrumental rather than cooperative rationality when success requires risk-bearing. The findings highlight framing effects as a significant bias source in non-interacting multi-agent LLM deployments, informing alignment and prompt design.

Figures & tables

$X_{i}$	$P_{B}$	Scenario A	Scenario B	Equivalence
A	$\geq 0.5$	$S_{i}=1$	$S_{\text{all}}=1$	$S_{i}=1$ (A-choosers)
A	$<0.5$	$S_{i}=1$	$S_{A}=1,S_{B}=0$	$S_{i}=1$ (A-choosers)
B	$\geq 0.5$	$S_{i}=1$	$S_{\text{all}}=1$	$S_{i}=1$ (B-choosers)
B	$<0.5$	$S_{i}=0$	$S_{A}=1,S_{B}=0$	$S_{i}=0$ (B-choosers)

Table 1: Logical Equivalence Verification: Scenario A vs. Scenario B

Panel of Figure 1 — Figure 1: Experimental workflow diagram

Family	Models	Scenario A A/B/C	$P(B)$	Scenario B A/B/C	$P(B)$	$\Delta P(B)$
Claude	14	210 / 0 / 70	0.000	0 / 200 / 80	0.714	0.714
DeepSeek	3	45 / 15 / 0	0.250	39 / 21 / 0	0.350	0.100
Doubao	4	80 / 0 / 0	0.000	13 / 67 / 0	0.838	0.838
GLM	8	152 / 7 / 1	0.044	52 / 74 / 34	0.463	0.419
GPT	18	322 / 32 / 6	0.089	73 / 256 / 31	0.711	0.622
Gemini	4	69 / 0 / 11	0.000	40 / 30 / 10	0.375	0.375

Table 2: Family-Level Response Summary Under Scenario A and Scenario B

Figure 2: Family-level response composition under Scenario A and Scenario B. The stacked bars show the proportions of A, B, and C responses for each family.

Figure 3: Framing effect magnitude $(\Delta P)$ by family. Positive values indicate higher preference for Option B under Scenario B.

Figure 4: Category C rate by family and prompt. Refusals are not the dominant outcome, but they are concentrated in a few families and are more visible in Scenario A for some models.

LLM Family	Test	Statistic	P-value	Significant
Claude	B vs. not-B	308.01	$<0.001$	Yes
DeepSeek	B vs. not-B	0.99	0.319	No
Doubao	B vs. not-B	111.85	$<0.001$	Yes
GLM	B vs. not-B	72.00	$<0.001$	Yes
GPT	B vs. not-B	287.78	$<0.001$	Yes
Gemini	B vs. not-B	34.50	$<0.001$	Yes

Table 3: Statistical Tests for Independence Between Framing and Choice Distribution by LLM Family

Figure 5: Exploratory open-CoT ablation using Has_Thinking . Bars show Option B probability under Scenario A and Scenario B for thinking-off and thinking-on subsets; sample sizes are annotated above bars.

Framing Effects in Language Models

Framing effects in language models study how wording, context, perspective, and presentation alter model judgments and outputs.

See all 51 papers

Framing Matters: Addressing Framing Sensitivity in Decision-Making through Behaviorally-Grounded Value Alignment
May 27, 2026Seojin Hwang, Minju Kim, Junhyuk Choi +2LLM Alignment Framing Effects in Language Models
Chung-Ang University, Seoul, Korea
Confirming Our Biases? Evaluating the Capabilities, Risks, and Societal Impact of Large Language Models
Aug 7, 2026Mudar Adas, Polina Tsvilodub, Michael Franke +1Prompt Sensitivity LLM Sycophancy
Neuro-Cognitive Modeling Group, University of Tübingen, Sand 14, 72076 Tübingen, Germany · Department of Linguistics, University of Tübingen, Keplerstraße 2, 72074 Tübingen
Auditing Framing-Sensitive Behavioral Instability in Large Language Models for Mental Health Interactions
Jun 25, 2026Abla Bedoui, Ashley L. Greene, Mohammed CherkaouiPrompt Sensitivity LLM Auditing
School of computer science, digital engineering and AI, Long Island University, Brooklyn, NY, USA · Department of Psychology, Long Island University, Brooklyn, NY, USA

May 27, 2026J/K move · Enter open · S save

Framing Matters: Addressing Framing Sensitivity in Decision-Making through Behaviorally-Grounded Value Alignment

Seojin Hwang, Minju Kim, Junhyuk Choi, JeongHyun Park+1

Chung-Ang University, Seoul, Korea

LLM Alignment Framing Effects in Language Models LLM Decision-Making Causal Interventions in Language Models Language Model Robustness

Open paper AskPDF arXiv

Large Language Models (LLMs) are increasingly deployed in high-stakes decision-making settings such as legal reasoning, where consistency under factually equivalent inputs is critical. However, we find that fact-preserved but differently framed inputs can significantly destabilize LLM decisions. To systematically investigate this problem, we introduce Fragile, a large-scale benchmark that isolates fact-preserving semantic framing across three controlled dimensions: value-tinted narration, temporal slice, and narrative vividness. Our experiments reveal a high susceptibility of LLMs to framing, with an average decision flip rate of 28.6%. We find that simple prior prompt-level and activation-level interventions not only fail to suppress framing sensitivity but actively amplify it. We therefore propose Valign, a representation-level method that explicitly targets these framing dimensions by anchoring decisions to a stable value prior, steering hidden states toward the model's value-consistent direction, and projecting out temporal-vividness-sensitive directions from the model's hidden states. Valign consistently reduces framing-induced decision flips, demonstrating that robust mitigation requires directly targeting the internal pathways in which framing operates.

Abstract

Framing Effects in Language Models

Explore similar work

Framing Matters: Addressing Framing Sensitivity in Decision-Making through Behaviorally-Grounded Value Alignment

Confirming Our Biases? Evaluating the Capabilities, Risks, and Societal Impact of Large Language Models

Auditing Framing-Sensitive Behavioral Instability in Large Language Models for Mental Health Interactions