Papers › Why We Feel: Breaking Boundaries in Emotional Reasoning with Multimodal Large Language Models

Why We Feel: Breaking Boundaries in Emotional Reasoning with Multimodal Large Language Models

10 Apr 2025arXiv:2504.07521archive 2025-07-28

Yuxiang Lin, Jingdong Sun, Zhi-Qi Cheng, Jue Wang, Haomin Liang, Zebang Cheng, Yifei Dong, Jun-Yan He, Xiaojiang Peng, Xian-Sheng Hua

Most existing emotion analysis emphasizes which emotion arises (e.g., happy, sad, angry) but neglects the deeper why. We propose Emotion Interpretation (EI), focusing on causal factors-whether explicit (e.g., observable objects, interpersonal interactions) or implicit (e.g., cultural context, off-screen events)-that drive emotional responses. Unlike traditional emotion recognition, EI tasks require reasoning about triggers instead of mere labeling. To facilitate EI research, we present EIBench, a large-scale benchmark encompassing 1,615 basic EI samples and 50 complex EI samples featuring multifaceted emotions. Each instance demands rationale-based explanations rather than straightforward categorization. We further propose a Coarse-to-Fine Self-Ask (CFSA) annotation pipeline, which guides Vision-Language Models (VLLMs) through iterative question-answer rounds to yield high-quality labels at scale. Extensive evaluations on open-source and proprietary large language models under four experimental settings reveal consistent performance gaps-especially for more intricate scenarios-underscoring EI's potential to enrich empathetic, context-aware AI applications. Our benchmark and methods are publicly available at: https://github.com/Lum1104/EIBench, offering a foundation for advanced multimodal causal analysis and next-generation affective computing.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2504.07521")

Code

Syntology Ran 5 of 9 code samples harvested from 1 repository linked to this paper; 4 have no recorded run. Of those that ran: 1 ran · our draft was wrong; 4 ran with no contract checked.

By repository: official repository: 9 samples from 1 repository, 5 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

lum1104/eibench officialmentioned in papermentioned on GitHubpytorchApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

9 samples harvested; 5 ran; 0 honoured the contract we drafted; 4 have no recorded run. Read from Syntology's graph 2026-09-25; that is when this build read the record, not when the samples ran.

1ran · our draft was wrong
4ran
4unverified

Licence: 0 of the 9 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from lum1104/eibench. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

calculate_average_score lum1104/eibench/EIBench/EI_Basic/get_scores.py official repository ran fingerprinted Apache-2.0 (permissive) · 3842caa4b493f017 · report
encode_image lum1104/eibench/EIBench/baselines/ChatGPT-4/gpt4-score-complex.py official repository ran · our draft was wrong Apache-2.0 (permissive) · f41cb1a19b154297 · report
extract_scores_from_jsonl lum1104/eibench/EIBench/EI_Basic/get_scores.py official repository ran Apache-2.0 (permissive) · 03387f6b6c21f170 · report
extract_scores_from_jsonl lum1104/eibench/EIBench/EI_Complex/get_scores_complex.py official repository ran Apache-2.0 (permissive) · 4807f06cfd55795c · report
get_ann lum1104/eibench/EIBench/human_eval/web_ann_basic.py official repository ran Apache-2.0 (permissive) · c1a40518b7d33c75 · report
ask_chatgpt lum1104/eibench/EIBench/EI_Basic/gpt-eval.py official repository unverified Apache-2.0 (permissive) · d12384275cefed30 · report
ask_chatgpt lum1104/eibench/EIBench/baselines/ChatGPT-4/gpt4-score-complex.py official repository unverified Apache-2.0 (permissive) · 6b60d87b77a0ce1c · report
load_jsonl lum1104/eibench/EIBench/human_eval/web_ann_basic.py official repository unverified Apache-2.0 (permissive) · e85a5515b1b5db8b · report
switch_label_based_on_dropdown lum1104/eibench/EIBench/human_eval/web_ann_basic.py official repository unverified Apache-2.0 (permissive) · e2378b5484a0e5e6 · report

Tasks

Emotion InterpretationEmotion Recognition

Datasets

Introduced by this paper, per the archive.

EIBench

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Emotion Interpretation EIBench Claude-3-haiku Recall 63.24 #1 of 13 Archive leaderboard report
Emotion Interpretation EIBench LLaVA-1.5 (13B) Recall 54.37 #2 of 13 Archive leaderboard report
Emotion Interpretation EIBench LLaVA-NEXT (13B) Recall 54.33 #3 of 13 Archive leaderboard report
Emotion Interpretation EIBench Claude-3-sonnet Recall 54.1 #4 of 13 Archive leaderboard report
Emotion Interpretation EIBench LLaVA-NEXT (7B) Recall 53.82 #5 of 13 Archive leaderboard report
Emotion Interpretation EIBench MiniGPT-v2 Recall 52.89 #6 of 13 Archive leaderboard report
Emotion Interpretation EIBench ChatGPT-4o Recall 49.99 #7 of 13 Archive leaderboard report
Emotion Interpretation EIBench Video-LLaVA Recall 49.26 #8 of 13 Archive leaderboard report
Emotion Interpretation EIBench LLaVA-NEXT (34B) Recall 49.03 #9 of 13 Archive leaderboard report
Emotion Interpretation EIBench ChatGPT-4V Recall 46.86 #10 of 13 Archive leaderboard report
Emotion Interpretation EIBench Otter Recall 42.81 #11 of 13 Archive leaderboard report
Emotion Interpretation EIBench Qwen-vl-plus Recall 31 #12 of 13 Archive leaderboard report
Emotion Interpretation EIBench Qwen-VL-Chat Recall 26.45 #13 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) ChatGPT-4o Recall 39.27 #1 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) LLaVA-NEXT (13B) Recall 39.16 #2 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) LLaVA-NEXT (7B) Recall 38.71 #3 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) LLaVA-1.5 (13B) Recall 38.1 #4 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) LLaVA-NEXT (34B) Recall 35.37 #5 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) MiniGPT-v2 Recall 35.1 #6 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) Video-LLaVA Recall 30.9 #7 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) ChatGPT-4V Recall 28 #8 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) Otter Recall 27.9 #9 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) Claude-3-haiku Recall 24 #10 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) Qwen-VL-Chat Recall 22 #11 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) Claude-3-sonnet Recall 21.37 #12 of 13 Archive leaderboard report
Emotion Interpretation EIBench (complex) Qwen-vl-plus Recall 20.37 #13 of 13 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections