Dataset Signatures in Human-LLM Interactions and User Modeling

Joseph Suh; Serina Chang

cs.CLOct 4, 2026

Dataset Signatures in Human-LLM Interactions and User Modeling

Authors: Joseph Suh, Serina Chang

Organizations: University of California, Berkeley

Abstract

Human--LLM interaction datasets shape our understanding of AI use and provide a foundation for downstream research, including training and evaluation of user models. In recent years, a growing number of datasets have sought to capture a representative picture of human--LLM interactions. But how different are the pictures these datasets provide, and what do those differences mean for research built on them? We study these questions across seven conversation datasets, spanning in-the-wild chat logs and human preference data. We begin by revisiting the dataset classification experiment of Torralba & Efros and find that neural network classifiers identify the source of a conversation from user messages alone well above chance, indicating distinctive dataset signatures. This separability persists after matching datasets on the dimensions of human-designed taxonomies, implying subtle differences that these taxonomies do not capture. We then examine the implications for user modeling: how dataset signatures propagate to the outputs of user models trained on these datasets; how dataset choice influences evaluations of user model quality and subsequent evaluations of LLM assistants paired with these user models; and how dataset classifiers can guide data selection for training user models. While each dataset is meant to capture a slice of 'real-world' interactions, our findings reveal the extent to which these slices diverge, and the consequences of those differences for research built on these foundations.

Figures & tables

Figure 1: Can you identify the source dataset? Nine example conversations, three each from WildChat-1M ( Zhao et al., 2024 ) , LMSYS-Chat-1M ( Zheng et al., 2024 ) , and HH-RLHF ( Bai et al., 2022 ) . On this three-way task, a neural classifier identifies the source with 83% overall accuracy and assigns over 90% probability to the correct source for each example shown here. Answers: WC-1M: 2, 6, 8; LMSYS: 3, 4, 9; HH-RLHF: 1, 5, 7.

Dataset	Description	Markers (prevalence)
WC-1M ( Zhao et al., 2024 )	WildChat-1M: voluntary interactions with free GPT-based services on HuggingFace Spaces.	PII redaction placeholder (0.8%)
WC-4.8M ( Zhao et al., 2024 )	An expanded release of WC-1M with newer assistant models, such as OpenAI o1.	PII redaction placeholder (1.3%)
LMSYS ( Zheng et al., 2024 )	LMSYS-Chat-1M: voluntary interactions on the Vicuna demo and Chatbot Arena platform.	NAME_ * (28.2%); [your answer] (1.1%)
Arena-human-preference-140K	Conversations reconstructed from released preference-vote histories during 2025.	-
ShareChat ( Yan et al., 2025 )	Publicly shared conversations retrieved from chatbot platforms.	<REDACTED> , <URL> , <DATE_TIME> (50.2%)
ShareGPT	User-shared ChatGPT history from ShareGPT.com. We use ShareGPT_Vicuna_unfiltered snapshot.	-

Table 1: Datasets used in our experiments. Please refer to Appendix A for filtering and sampling details. Marker prevalence is the percentage of conversations containing dataset-specific markers.

Dataset

\downarrow

Setting

\rightarrow

Binary

Three-way

Incremental

WC-1M

\checkmark

–

\checkmark

\checkmark

\checkmark

–

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

WC-4.8M

\checkmark

–

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

LMSYS

–

\checkmark

–

\checkmark

\checkmark

–

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

\checkmark

Arena

–

\checkmark

–

\checkmark

\checkmark

\checkmark

\checkmark

ShareChat

–

\checkmark

\checkmark

–

\checkmark

–

\checkmark

\checkmark

\checkmark

ShareGPT

–

\checkmark

–

\checkmark

–

\checkmark

\checkmark

Table 2: Dataset classification shows test accuracy well above chance. For each dataset combination, we train a multinomial logistic regression head that takes Qwen2.5-7B representations as an input.

Figure 3: Accuracy on WC-1M + LMSYS + ShareChat across training sample sizes and model sizes.

	Accuracy (%)
Pair	Original	F	F+T	F+T+M	F+T+M+S	$\Delta$ [pp]
WC-1M / LMSYS	78.7	77.3	75.4	75.4	74.9	$-$ 3.8
WC-1M / ShareGPT	86.2	83.5	82.9	83.8	83.2	$-$ 3.0
LMSYS / ShareGPT	83.4	80.3	79.9	77.2	77.4	$-$ 6.0

Table 4: Classification accuracy (%) after matching category distributions (chance is 50%). From left to right, columns add matched facets cumulatively: F (Function), T (Topic), M (Multi-turn), and S (Structural: numbers of turns and user tokens).

\Delta

is the change from Original to F+T+M+S.

Figure 4: Category differences before and after matching on classifier scores. For each dataset pair and taxonomy facet, hollow and filled circles show the TVD between category distributions before and after matching; the bar marks the noise floor, the median TVD between two random halves of the same dataset across 1,000 bootstrap samples. Matching moves distances toward the noise floor.

Dataset pair	Real $\to$ real	Intent source	Real $\to$ synth.	Synth. $\to$ synth.
WC-1M / WC-4.8M	62.8	HH-RLHF	59.3	62.9
WC-4.8M / LMSYS	85.1	WC-1M	67.4	74.5
WC-4.8M / ShareChat	90.3	WC-1M	73.7	79.3
LMSYS / ShareChat	88.7	WC-4.8M	70.0	75.8
LMSYS / ShareGPT	83.4	WC-4.8M	63.6	72.8
ShareChat / ShareGPT	86.8	LMSYS	69.7	74.6

Table 5: Dataset classification accuracy (%) on synthetic conversations with Llama-based user models (chance is 50%). For each pair, two user models are conditioned on the same intents from the intent source. Real

\to

real is a reference accuracy, training and testing on real conversations. Real

\to

synth.: train on real, test on synthetic. Synth.

\to

synth.: train and test on synthetic.

	GSM8K			HumanEval
User model	Qwen	Llama	$\Delta$	Qwen	Llama	$\Delta$
WC-1M	9.6 $\pm$ 2.6	10.0 $\pm$ 3.1	-0.4 $\pm$ 2.6	14.2 $\pm$ 3.2	8.4 $\pm$ 2.6	5.8 $\pm$ 3.1
WC-4.8M	16.2 $\pm$ 4.1	13.8 $\pm$ 3.4	2.4 $\pm$ 3.9	18.0 $\pm$ 3.8	11.0 $\pm$ 3.2	7.0 $\pm$ 2.6
LMSYS	18.2 $\pm$ 4.2	17.6 $\pm$ 3.5	0.6 $\pm$ 3.7	18.0 $\pm$ 4.2	14.0 $\pm$ 3.7	4.0 $\pm$ 3.8
ShareChat	26.8 $\pm$ 4.4	22.0 $\pm$ 4.5	4.8 $\pm$ 4.5	17.0 $\pm$ 4.0	12.0 $\pm$ 3.2	5.0 $\pm$ 3.1
ShareGPT	24.8 $\pm$ 3.7	14.4 $\pm$ 2.9	10.4 $\pm$ 4.0	21.2 $\pm$ 4.6	9.8 $\pm$ 3.1	11.4 $\pm$ 4.3

Table 6: Task success rate (%) of Qwen2.5-14B-Instruct (Qwen) and Llama-3.1-8B-Instruct (Llama) as assistants when interacting with user models trained on different datasets (rows). Each entry averages 500 simulated conversations (100 problems

\times

5). 90% CI (

\pm

) is calculated by first averaging the five outcomes for each problem, then using a Student’s t interval across the problem-level averages.

Train $\downarrow$ Test $\rightarrow$	WC-1M	WC-4.8M	LMSYS	Arena	ShareChat	ShareGPT	HH-RLHF
WC-1M	6.41	6.70	6.14	6.15	8.38	6.02	5.39
WC-4.8M	6.53	6.36	6.17	6.06	8.28	6.07	5.46
LMSYS	6.61	6.80	5.43	6.17	8.31	6.03	5.19
Arena	6.69	6.82	6.22	5.75	8.29	6.14	5.29
ShareChat	6.71	6.85	6.22	6.13	7.33	6.09	5.30
ShareGPT	6.65	6.92	6.21	6.26	8.40	5.62	5.28

Table 7: Cross-dataset evaluation. Rows indicate training and columns indicate test (target) datasets. Each entry represents mean per-token perplexity measured on the test split. Bold and underlined entries indicate the lowest and second-lowest perplexity in each column, respectively.

Figure 5: Symmetric transfer gap

\Delta_{AB}

correlates with the

A

-to-

B

dataset classification accuracy.

Figure 6: Perplexity reduction relative to uniform random sampling at 50% of the training budget, PPL(50% random) - PPL(condition), for Qwen2.5-7B-Instruct user models (higher is better). The number under each dataset is the 50%-random perplexity. Classifier-guided selection outperforms random sampling at every matched budget. Llama-3-8B results are in Figure 8 .

Appendix figures & tables8 assets

Supplementary material from the paper’s appendix.

Appendix

Dataset	Input	English	Eligible	Retained
HH-RLHF	169,352	168,805	145,740	50,000
ShareGPT	76,920	54,352	50,790	50,000
LMSYS-Chat-1M	1,000,000	766,545	415,087	50,000
Arena 140K (2025)	135,634	60,810	57,184	50,000
WildChat-1M	837,989	473,246	285,771	50,000
WildChat-4.8M	3,199,860	1,672,053	225,212	50,000

Appendix

Table 8: Curation and sampling counts for the snapshots used in this study. Input counts are raw records. English counts follow conversation reconstruction, structural validation, and language filtering. Eligible counts additionally reflect duplicate/template filtering and cross-source exclusions.

Figure 7: Percentage of conversations containing each annotated category in the original held-out WC-1M, LMSYS, and ShareGPT cohorts. A conversation can contain several categories, so columns need not sum to 100%. Each factor uses only conversations with complete accepted annotations.

Model size	Frozen	Fine-tuned	$\Delta$ [%]
0.5B	69.7	80.3	+10.6
1.5B	70.1	81.4	+11.3
3B	72.5	80.2	+7.7
7B	72.5	79.3	+6.8

Appendix

Table 9: Frozen language models vs. co-trained classifiers on WC-1M + LMSYS + ShareChat.

		Real $\to$ synth.			Synth. $\to$ synth.
Dataset pair	Intent source	Qwen3.5	Llama3.1	Qwen2.5	Qwen3.5	Llama3.1	Qwen2.5
WC-1M / WC-4.8M	HH-RLHF	59.3	58.8	59.2	62.9	62.6	64.2
WC-4.8M / LMSYS	WC-1M	67.4	67.5	66.9	74.5	72.9	74.3
WC-4.8M / ShareChat	WC-1M	73.7	73.9	74.0	79.3	78.8	78.4
LMSYS / ShareChat	WC-4.8M	70.0	70.6	70.3	75.8	76.4	77.3
LMSYS / ShareGPT	WC-4.8M	63.6	64.8	65.2	72.8	70.6	70.0

Appendix

Table 10: User-model classification accuracy (%) on synthetic conversations with three different assistants (chance is 50%). For each pair, two user models are conditioned on the same intents from the intent source and interact with the same assistant; the Qwen3.5-9B columns repeat Table 5 . Real

\to

synth.: train on real, test on synthetic. Synth.

\to

synth.: train and test on synthetic.

		User turns only			Assistant
Comparison	Chance	All	Single-turn	Multi-turn	turns only
Qwen3.5 vs. Llama3.1	50.0	53.4 (51.0 – 57.5)	50.5 (49.2 – 52.4)	54.6 (50.5 – 62.7)	99.2 (98.3 – 99.7)
Qwen3.5 vs. Qwen2.5	50.0	54.1 (50.7 – 59.6)	50.7 (49.4 – 52.0)	55.4 (50.9 – 63.8)	98.8 (97.8 – 99.6)
Qwen2.5 vs. Llama3.1	50.0	50.8 (49.8 – 53.0)	50.1 (49.4 – 51.1)	51.3 (50.0 – 54.9)	98.8 (97.8 – 99.6)
Three-way	33.3	35.9 (33.6 – 38.8)	33.7 (32.9 – 34.6)	36.6 (33.6 – 42.1)	98.3 (96.7 – 99.4)

Appendix

Table 11: Assistant classification accuracy (%) on synthetic conversations with the same intents and user model. Entries are means over 17 combinations of intent source and user model, with the range in parentheses. The first three columns use user turns only, with assistant responses masked; single-turn conversations have one user message and multi-turn conversations two or more. The last column uses assistant responses only, with user turns masked.

Train $\downarrow$ Test $\rightarrow$	WC-1M	WC-4.8M	LMSYS	Arena	ShareChat	ShareGPT	HH-RLHF
WC-1M	6.60	7.05	6.72	6.65	8.89	6.70	5.78
WC-4.8M	6.78	6.47	6.81	6.54	8.75	6.77	5.84
LMSYS	6.93	7.30	5.60	6.71	8.90	6.72	5.60
Arena	7.03	7.33	6.91	6.10	8.93	6.93	5.76
ShareChat	7.10	7.36	6.94	6.71	7.70	6.85	5.83
ShareGPT	7.04	7.44	6.79	6.82	8.90	5.93	5.62

Appendix

Table 12: Cross-dataset evaluation. Rows indicate training and columns indicate test (target) datasets. Each entry represents mean per-token perplexity measured on the test split. Bold and underlined entries indicate the lowest and second-lowest perplexity in each column, respectively.

Figure 8: Perplexity reduction relative to uniform random sampling at 50% of the training budget, PPL(50% random) - PPL(condition), for Llama-3-8B user models (higher is better). The number under each dataset is the 50%-random perplexity. Classifier-guided selection outperforms random sampling at every matched budget.

Figure 9: Magnitude of classifier-score distribution change versus downstream NLL improvement. The horizontal axis shows Equation 4 , computed from trained classifier scores for target, randomly selected donor, and classifier-selected donor. The vertical axis shows target NLL reduction,

\Delta_{\text{NLL}}

Explore similar work

May 27, 2026cs.AI

Adopt $\neq$ Adapt: Longitudinal Analyses of LLM Conversations in the Wild

Although a growing body of research has begun to describe user--LLM interactions, the picture it paints is largely static; little is known about how individual users change their behavior over time. To address this gap, we analyze the conversational trajectories of ~12,000 randomly sampled Microsoft Bing Copilot users and compare these with data from WildChat-4.8M. While the Copilot data contains significant population-level trends, we find that trends in individual user trajectories are much weaker; user habits prove to be overwhelmingly sticky. We also find stark differences between users of different activity levels: more active users have more successful conversations and use the LLM for more complex and professionally oriented tasks. Some user trends also appear in WildChat-4.8M, but we find evidence that this dataset is significantly skewed towards highly proficient "power" users. Ultimately, our results suggest that existing user behavior is difficult to change and demonstrate the extent of user heterogeneity. Our comparison between datasets highlights that WildChat does not represent typical user--AI interactions, an important caveat for downstream uses of the data.

Rebecca M. M. Hicke, Kiran Tomlinson

Cornell University · Microsoft Research

Multi-Turn Large Language Model Conversations Copilot Longitudinal Data Heterogeneity

Ask PDF

Jun 4, 2026cs.CL

Re-Centering Humans in LLM Personalization

Despite growing interest, most evaluations of large language models' (LLMs') personalization abilities have relied on synthetic data. It remains unclear how well current personalization systems work for real users. In this paper, we study the gap in LLM personalization performance when using synthetic versus human data. We collect human conversations (550 conversations) and judgments across three stages of personalization: extracting user attributes from conversations (5,949 judgments), pairing relevant attributes with new prompts (11,919), and incorporating relevant attributes into a personalized response (1,101). Incorporating human data reveals system limitations at each stage. Models struggle to extract attributes from human conversations, disagree with human judgments on relevant attributes, and generate personalized responses that humans judge no better than generic responses (though that LLM judges widely rate as better). We introduce two lightweight training-based interventions that shift automated personalization evaluation closer to human data in our first two stages. However, in our third stage we find that learned reward models achieve only modest correlation with human ratings, suggesting that human-aligned personalization quality judgments are difficult to model directly. Our collected data provides a foundation for studying how models should extract, select, and incorporate user information in ways that humans find useful.

Lechen Zhang, Jiarui Liu, Tal August

†University of Illinois Urbana-Champaign · ♯Carnegie Mellon University

Large Language Model Personalization Personalization Large Language Model Evaluation Human Agency Synthetic Data

Ask PDF

May 19, 2026cs.CL

ThoughtTrace: Understanding User Thoughts in Real-World LLM Interactions

Conversational AI has now reached billions of users, yet existing datasets capture only what people say, not what they think. We introduce ThoughtTrace, the first large-scale dataset that pairs real-world multi-turn human--AI conversations with users' self-reported thoughts: their reasons for sending prompts and reactions to assistant responses. ThoughtTrace comprises 1,058 users, 2,155 conversations, 17,058 turns, and 10,174 thought annotations collected across 20 language models. Our analysis shows that ThoughtTrace captures long-horizon, topically diverse interactions, and that thoughts are semantically distinct from messages, difficult for frontier LLMs to infer from context, diverse in content, and tied to conversation stages. We further demonstrate the utility of thoughts for downstream modeling. First, thoughts improve user-behavior prediction as inference-time context. Second, thought-guided rewrites provide fine-grained alignment signals for training personalized assistants. Together, ThoughtTrace establishes user thoughts as a new data modality for studying the cognitive dynamics behind human--AI interaction and provides a foundation for building assistants that better understand and adapt to users' latent goals, preferences, and needs.

Chuanyang Jin, Binze Li, Haopeng Xie +6

Johns Hopkins University · Massachusetts Institute of Technology · Google Research

Thoughts Conversational Artificial Intelligence Cognitive Science Traces Annotations

Ask PDF

Abstract

Explore similar work

Adopt ≠\neq= Adapt: Longitudinal Analyses of LLM Conversations in the Wild

Re-Centering Humans in LLM Personalization

ThoughtTrace: Understanding User Thoughts in Real-World LLM Interactions

Adopt $\neq$ Adapt: Longitudinal Analyses of LLM Conversations in the Wild