PsyLite Technical Report

Chengye Xie; Fangjun Ding; Renyu Zhang; Xinyu Feng; Yanting Zhang; Zheng Zhang

REVIEW 4 major objections 4 minor 23 references

PsyLite Technical Report

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Fine-tuned 7B model lifts counseling professionalism by 47.6%.

desk verdict A useful, honest applied-LLM report whose headline counseling gain currently rests on evaluation evidence too weak to trust; the corrected claim is testable and the paper deserves a serious referee. read the letter →

arxiv 2506.21536 v1 pith:AIEYFHHK submitted 2025-06-26 cs.AI cs.HC

Fangjun Ding , Renyu Zhang , Xinyu Feng , Chengye Xie , Zheng Zhang , Yanting Zhang This is my paper

classification cs.AIcs.HC

keywords lightweightpsychologicalcounselingLLMdeepreasoningdialoguesafetydeploymentORPOpreferenceoptimizationQLoRAfine-tuningconditionalretrieval-augmentedgenerationcrosstalkhumor

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

PsyLite sets out to show that a 7B-parameter Chinese dialogue model can be turned into a credible psychological counseling assistant while staying light enough to run locally. The authors fine-tune InternLM2.5-7B-chat in two stages: first a QLoRA supervised pass on a 10:3 mix of general chain-of-thought distillation data and counseling dialogues with chain-of-thought injected, then ORPO preference optimization on safe-dialogue data. On the CPsyCounE counseling evaluation they report professionalism rising from 1.89 to 2.83, a 47.6% gain, and on SafeDialBench safety rising from 8.72 to 8.93, at the cost of a small CEval drop from 78.07 to 76.56. After q4_k_m quantization, they claim the deployed agent runs in about 5GB and uses a conditional retrieval workflow to insert crosstalk humor only when the user is in a good mood and to refuse dangerous requests. If these numbers hold, a professional-sounding, safety-conscious counseling assistant no longer requires a large GPU cluster.

What carries the argument

The load-bearing mechanism is the hybrid distillation dataset psy-mix-gen-distill-13k, built as 10k general chain-of-thought distillation samples plus 3k psychological counseling dialogues that were first run through a strong reasoning model to inject chain-of-thought between the client's question and the counselor's answer. Training this mixture in one QLoRA supervised fine-tuning pass, rather than in separate domain and reasoning steps, is what the authors say preserves general ability while adding counseling professionalism; the 10:3 ratio is their designed balance point. The second mechanism is ORPO, a preference-optimization objective that combines supervised loss with an odds-ratio loss favoring the chosen safe response over the rejected one, which the paper credits for the safer refusal behavior. In the deployed agent a conditional RAG workflow sits on top: an inlet filter classifies the user's state, and only in a pleasant state does a retrieval step pull crosstalk humor, while dangerous states trigger preset refusal phrases.

What would settle it

Have independent raters who are blind to model identity and do not know the training data re-score the same CPsyCounE prompts; if the fine-tuned model's professionalism margin over the base model collapses below a few percent, the claimed 47.6% improvement is not a real counseling-quality gain. A second check would build a fresh counseling test set from new therapy transcripts unrelated to CPsyCoun and see whether the margin survives.

Watch

Extended reading notes

Core claim

The paper's central claim is that one two-stage training recipe is enough to specialize a general 7B chat model for psychological counseling without sacrificing its general abilities or its refusal behavior. Stage one mixes 10,000 general reasoning-distillation samples with 3,000 counseling dialogues that have been enriched with chain-of-thought explanations, and trains the mixture in a single QLoRA supervised fine-tune; stage two applies ORPO on the PKU-SafeRLHF single-dimension preference set to make rejections of unsafe requests more likely. The authors report that the resulting InternLM2.5-7B-distill-orpo beats the base InternLM2.5-7B-chat on CPsyCounE (professionalism 2.83 vs 1.89, comprehensiveness 1.97 vs 1.76, authenticity 2.72 vs 2.52, safety 1.00 vs 1.00) and on SafeDialBench (8.93 vs 8.72), and remains close on CEval (76.56 vs 78.07). They further claim the quantized GGUF model plus a Pipelines-based conditional RAG workflow, which adds crosstalk humor in pleasant states and blocks dangerous requests, gives a usable counseling agent on about 5GB of memory.

Load-bearing premise

The headline professionalism gain assumes the authors' own CPsyCounE manual scoring is an unbiased measure of counseling quality, even though the evaluation prompts and the training data come from the same report-reconstruction framework and no inter-rater reliability or blinding is reported.

Editorial extensions

If this is right

If the reported gains hold, a 7B model fine-tuned with QLoRA and ORPO is enough to lift counseling professionalism scores by nearly half, so specialized mental-health models no longer require multi-billion-parameter training runs.
The 10:3 hybrid mixing ratio with chain-of-thought injection becomes a reusable recipe for other verticals that need both deep reasoning and domain knowledge.
The modest SafeDialBench gain (8.93 vs 8.72) implies ORPO on a preference set adds a real but limited safety margin on top of the base model's own refusal behavior.
Quantizing to GGUF q4_k_m makes the trained model run in around 5GB, which would allow fully offline counseling support on a laptop or edge device with no data leaving the user's machine.
The conditional RAG design shows humor and safety can coexist by gating retrieval on a state classification rather than baking both into the model's weights.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The 47.6% CPsyCounE improvement may partly reflect distribution overlap, since the training dialogues and the evaluation prompts come from the same CPsyCoun report-reconstruction framework; an out-of-family counseling test set would be needed to confirm the gain is general.
Manual scoring by the authors, without reported blinding or inter-rater agreement, is the weakest link; a blinded multi-rater replication would probably compress the reported gap.
Because the evaluation used only the 249 available jailbreak samples, a larger safety test across more attack strategies would clarify whether the +0.21 SafeDialBench gain is robust.
The paper reports no ablation separating the general distillation data, the CoT-injected counseling data, and the ORPO stage; such an ablation would identify which ingredient drives the professionalism gain.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

You should know this before opening the PDF: it's a straightforward applied LLM technical report, not a methods breakthrough. The components are all prior art — QLoRA SFT, R1 distillation, ORPO, RAG, GGUF quantization. What's actually new is the specific 10:3 general-to-counseling distilled data mix, the CoT injection into CPsyCounD using DeepSeek R1, and a conditional RAG that retrieves crosstalk humor when the user is in a good mood. Those are real, concrete artifacts, and the paper is refreshingly honest in places: the body explicitly admits the CEval score is lower than baseline, and the failed-attempts section describes ORPO hyperparameter struggles without spin.

The soft spots are exactly where the reader's report points. The abstract claims CEval outperformance, but Table 1 shows 76.56 vs 78.07 — a decline. More importantly, the load-bearing CPsyCounE result (2.83 vs 1.89, the 47.6% improvement) comes from manual scoring by the authors, with no sample size, no inter-rater reliability, no blinding, and no variance. Since the training data CPsyCounD and the eval set CPsyCounE both come from the same report-based reconstruction framework, the gain could partly reflect distribution matching rather than genuine counseling skill. The SafeDialBench result is more modest and rests on 249 samples with DeepSeek R1 auto-scoring plus manual review. The ORPO checkpoint was selected on training reward, which is cherry-picking of a sort.

That said, these are fixable flaws, not fatal ones. The paper doesn't hide the CEval decline; it argues the small general-knowledge cost is acceptable. The central claim — a 7B model can improve counseling professionalism and safety at a small general-task cost — is plausible and testable. What's missing is rigor: independent blind rescoring of CPsyCounE, a stated sample size and reliability statistic, and ideally a held-out eval set not derived from the same framework as the training data. Releasing the model and data with a commit hash would also help.

Who gets value from this? Researchers working on lightweight mental-health LLMs or on evaluation pitfalls in domain-specific fine-tuning. It's a good case study for a reading group on why benchmark overlap and manual scoring matter. I'd send it to peer review, not desk-reject it — the work is honest and the recipe is reproducible — but I'd make the revision contingent on tightening the evaluation. If I were refereeing, my main request would be: rerun the CPsyCounE scoring with an independent, blinded rater or two, report inter-rater agreement, and either fix the abstract or drop the CEval outperformance claim.

Referee Report

4 major / 4 minor

Summary. The paper describes PsyLite, a lightweight Chinese psychological counseling LLM built on InternLM2.5-7B-chat. Training is two-stage: first, supervised fine-tuning with QLoRA on a mixed dataset of 10k general DeepSeek-R1 distilled items and 3k psychological counseling items (CPsyCounD with injected chain-of-thought); second, ORPO preference optimization on PKU-SafeRLHF to improve dialogue safety. The system is then quantized to GGUF q4_k_m and deployed via Ollama and Open WebUI, with a Pipelines-based conditional RAG mechanism that inserts crosstalk humor when the user is in a positive mood and refuses dangerous requests. The authors report evaluations on CEval, CPsyCounE, and SafeDialBench, claiming strong gains in psychological counseling professionalism (47.6%) and dialogue safety (2.4%), while maintaining a 5GB deployment footprint.

Significance. If the reported gains are real, the result is practically valuable: it demonstrates that a 7B open model can be adapted to psychological counseling with meaningful improvements in professionalism and safety, and that the result can be deployed on low-resource hardware. The training pipeline is described in sufficient detail to be reproduced from public datasets, and the deployment choices (GGUF, Ollama, Open WebUI, Pipelines) are concrete and actionable. However, the central evidence is weakened by evaluation methodology: the CPsyCounE scores are manually assigned by the authors without blinding, inter-rater reliability, variance, or sample-size reporting; the evaluation set is drawn from the same reconstruction framework as the training data; and the safety result rests on a small 249-sample jailbreak subset with auto-scoring. The abstract also claims CEval outperformance that the body explicitly contradicts.

major comments (4)

[§4.2.2, Table 2, Abstract] The headline 47.6% improvement in CPsyCounE professionalism (2.83 vs 1.89) rests on manual scoring by the authors with no reported number of evaluated dialogues, no item-level scores, no variance or confidence interval, and no inter-rater reliability or blinding. Because the scorers know which outputs come from the fine-tuned model, the large gap could reflect expectation or style preference rather than measurable counseling professionalism. This is the load-bearing claim of the paper, and I do not see sufficient evidence for it. Please have independent, blinded raters rescore the outputs, report the scoring protocol and sample size, and provide inter-rater agreement statistics (e.g., Cohen's kappa or ICC).
[§3.1.1, §2.2.1, §4.2.2] There is a data-dependence concern between training and evaluation: the training set CPsyCounD and the evaluation benchmark CPsyCounE are both products of the same report-based reconstruction framework described in [Zhang et al., 2024]. The model may therefore be rewarded for matching the style and distribution of that framework rather than for generalizable counseling skill. The paper should either evaluate on an independent set of human-annotated counseling dialogues or explicitly quantify the overlap and show that the gain persists on examples generated outside the CPsyCoun reconstruction pipeline.
[§4.2.3, Table 3] The SafeDialBench result (8.93 vs 8.72) is based on only 249 jailbreak samples, with scores from DeepSeek R1 auto-scoring plus manual review. The paper does not report how many outputs were manually reviewed, what the manual review protocol was, the variance of the scores, or whether the 0.21 difference is statistically significant. With 249 samples, this difference may well be within noise. Please report additional statistics, clarify the manual-review contribution, and state whether the subset of 249 is representative of the full 4,000+ conversation benchmark described in §2.2.2.
[Abstract, §4.2.1, Table 1] The abstract states that PsyLite 'outperforms the baseline models' on CEval, but Table 1 reports 76.56 for PsyLite versus 78.07 for the baseline InternLM2.5-7B-chat, and §4.2.1 explicitly concedes that the model scored slightly lower on CEval. This is a direct contradiction between the abstract and the body. The abstract must be revised to describe the CEval result accurately, or the claim must be withdrawn.

minor comments (4)

[§3.1.2, Eq. (3)] The phrase 'does not require training and rewards' is misleading; ORPO is a training objective and does not require a separate reward model, but it certainly requires training. Please rephrase to avoid confusion.
[§3.3.2, Figures 6 and 7] The text says 'Figure 7 is the Pipelines workflow' but the preceding figure is labeled Figure 6, and later text refers to 'the following Figure 7' for the usage case. The figure numbering appears to be off by one; please fix the cross-references.
[References] The reference for SoulChat in §2.3.3 is listed under the title 'Psydt: Using llms to construct the digital twin...' (Xie et al., 2024), which does not match the name SoulChat used in the text. Please verify the correct citation and ensure the reference title matches the cited work.
[§5.3] The failed-attempts section describes ORPO hyperparameter selection and says the best checkpoint was chosen after 3k steps at learning rate 5e-6 and beta=0.2, but it does not state how this checkpoint selection relates to the final reported SafeDialBench and CPsyCounE results. Please clarify the model-selection procedure and whether the reported results come from this checkpoint.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: reported gains are empirical evaluations with acknowledged limitations, not constructionally forced results.

full rationale

PsyLite's reported gains are empirical measurements from fine-tuned model evaluations, not quantities derived from a model whose definition already contains the measured value. The training stage fits QLoRA SFT and ORPO on CPsyCounD and PKU-SafeRLHF; the evaluation stage scores CEval, CPsyCounE, and SafeDialBench separately. No equation in the paper defines the reported CPsyCounE or SafeDialBench scores in terms of the training loss or the distillation data, so no result reduces to its inputs by construction. The closest issue is that CPsyCounD and CPsyCounE share the same CPsyCoun report-reconstruction framework (Sections 2.2.1, 3.1.1, and 4.2.2), which weakens the independence of the evaluation; however, the paper does not state that identical dialogues appear in both sets, and shared provenance is not the definitional equivalence required for circularity. The Section 5.3 admission that the ORPO checkpoint was selected at a local training-reward maximum is a model-selection limitation, not a fitted parameter renamed as a prediction. The manual, unblinded scoring of CPsyCounE is a validity concern but belongs to evaluation rigor, not circularity. No load-bearing self-citations or ansatz-smuggling citations are present, so the derivation chain is self-contained.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper rests on standard LLM training infrastructure plus several hand-chosen design parameters. The main unverified inputs are the validity of the manual CPsyCounE scoring, the reliability of the 249-sample SafeDialBench evaluation, the assumption that R1 distillation improves reasoning in a 7B model, and the assumption that ORPO safety training transfers to the SafeDialBench distribution. No new physical or mathematical entities are introduced.

free parameters (4)

Data mixing ratio (general:psychological) = 10:3
Section 3.1.1: chosen by hand to balance generalization and counseling ability; no objective criterion is given, and the ratio directly affects the reported CPsyCounE and CEval trade-off.
Distill data filtering score threshold = 9
Section 3.1.1: general distill samples are filtered to score >= 9 before random sampling; this threshold is arbitrary and not justified.
ORPO beta and checkpoint step = beta=0.2, step=3000
Sections 3.1.2 and 5.3: the checkpoint at 3k steps was selected because reward_margin and reward_acc reached local maxima; this is selection on training reward and affects the reported safety result.
SFT learning rate and epochs = 1e-4, 5 epochs
Section 3.1.1: training hyperparameters; part of the recipe but not fitted to a target, included for completeness.

assumptions (5)

domain assumption Distilling DeepSeek R1 outputs into a 7B model improves reasoning and downstream performance.
Section 3.1.1 cites the DeepSeek-R1 tech report; the paper assumes this transfer without running a reasoning-specific benchmark to verify the 'deep-reasoning ability' claim.
domain assumption ORPO on PKU-SafeRLHF improves dialogue safety without degrading other abilities.
Section 3.1.2 assumes the safety benefit; the paper reports only a 0.21-point SafeDialBench gain with no significance testing.
domain assumption CPsyCounE manual scoring is a valid measure of counseling professionalism.
Section 4.2.2 relies on author manual scoring; no inter-rater reliability or external validation is provided.
domain assumption The Inlet Filter can reliably classify user emotional state into dangerous, normal, or pleasant categories.
Section 3.3.2 depends on an LLM-based state assessment; no accuracy evaluation of this classifier is given.
standard math Standard autoregressive language modeling equations (Eqs. 1-2) hold for the fine-tuning objective.
Section 3.1.1 uses standard cross-entropy; this is unproblematic background math.

how reviews work

0 comments

Cite this review

Pith. "Pith review of PsyLite Technical Report." pith.science (2026). https://pith.science/paper/AIEYFHHK

@misc{pith2026250621536,
  author       = {Pith},
  title        = {Pith review of: PsyLite Technical Report},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AIEYFHHK}},
  note         = {Machine review of arXiv:2506.21536}
}

read the original abstract

With the rapid development of digital technology, AI-driven psychological counseling has gradually become an important research direction in the field of mental health. However, existing models still have deficiencies in dialogue safety, detailed scenario handling, and lightweight deployment. To address these issues, this study proposes PsyLite, a lightweight psychological counseling large language model agent developed based on the base model InternLM2.5-7B-chat. Through a two-stage training strategy (hybrid distillation data fine-tuning and ORPO preference optimization), PsyLite enhances the model's deep-reasoning ability, psychological counseling ability, and safe dialogue ability. After deployment using Ollama and Open WebUI, a custom workflow is created with Pipelines. An innovative conditional RAG is designed to introduce crosstalk humor elements at appropriate times during psychological counseling to enhance user experience and decline dangerous requests to strengthen dialogue safety. Evaluations show that PsyLite outperforms the baseline models in the Chinese general evaluation (CEval), psychological counseling professional evaluation (CPsyCounE), and dialogue safety evaluation (SafeDialBench), particularly in psychological counseling professionalism (CPsyCounE score improvement of 47.6\%) and dialogue safety (\safe{} score improvement of 2.4\%). Additionally, the model uses quantization technology (GGUF q4\_k\_m) to achieve low hardware deployment (5GB memory is sufficient for operation), providing a feasible solution for psychological counseling applications in resource-constrained environments.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 11 canonical work pages

[1]

J. Bai, S. Bai, Y. Chu, Z. Cui, K. Dang, X. Deng, Y. Fan, W. Ge, Y. Han, F. Huang, B. Hui, L. Ji, M. Li, J. Lin, R. Lin, D. Liu, G. Liu, C. Lu, K. Lu, J. Ma, R. Men, X. Ren, X. Ren, C. Tan, S. Tan, J. Tu, P. Wang, S. Wang, W. Wang, S. Wu, B. Xu, J. Xu, A. Yang, H. Yang, J. Yang, S. Yang, Y. Yao, B. Yu, H. Yuan, Z. Yuan, J. Zhang, X. Zhang, Y. Zhang, Z. Zh...

arXiv 2023
[2]

Z. Cai, M. Cao, H. Chen, K. Chen, K. Chen, X. Chen, X. Chen, Z. Chen, Z. Chen, P. Chu, X. Dong, H. Duan, Q. Fan, Z. Fei, Y. Gao, J. Ge, C. Gu, Y. Gu, T. Gui, A. Guo, Q. Guo, C. He, Y. Hu, T. Huang, T. Jiang, P. Jiao, Z. Jin, Z. Lei, J. Li, J. Li, L. Li, S. Li, W. Li, Y. Li, H. Liu, J. Liu, J. Hong, K. Liu, K. Liu, X. Liu, C. Lv, H. Lv, K. Lv, L. Ma, R. Ma...

arXiv 2024
[3]

H. Cao, Y. Wang, S. Jing, Z. Peng, Z. Bai, Z. Cao, M. Fang, F. Feng, B. Wang, J. Liu, T. Yang, J. Huo, Y. Gao, F. Meng, X. Yang, C. Deng, and J. Feng. Safedialbench: A fine-grained safety benchmark for large language models in multi-turn dialogues with diverse jailbreak attacks, 2025. URL https://arxiv.org/abs/2502.11090

arXiv 2025
[4]

DeepPsy-Agent: A Stage-Aware and Deep-Thinking Emotional Support Agent System

K. Chen and Z. Sun. Deeppsy-agent: A stage-aware and deep-thinking emotional support agent system, 2025. URL https://arxiv.org/abs/2503.15876

work page Pith review arXiv 2025
[5]

DeepSeek-AI, :, X. Bi, D. Chen, G. Chen, S. Chen, D. Dai, C. Deng, H. Ding, K. Dong, Q. Du, Z. Fu, H. Gao, K. Gao, W. Gao, R. Ge, K. Guan, D. Guo, J. Guo, G. Hao, Z. Hao, Y. He, W. Hu, P. Huang, E. Li, G. Li, J. Li, Y. Li, Y. K. Li, W. Liang, F. Lin, A. X. Liu, B. Liu, W. Liu, X. Liu, X. Liu, Y. Liu, H. Lu, S. Lu, F. Luo, S. Ma, X. Nie, T. Pei, Y. Piao, J...

arXiv 2024
[6]

DeepSeek-AI, D. Guo, D. Yang, H. Zhang, J. Song, R. Zhang, R. Xu, Q. Zhu, S. Ma, P. Wang, X. Bi, X. Zhang, X. Yu, Y. Wu, Z. F. Wu, Z. Gou, Z. Shao, Z. Li, Z. Gao, A. Liu, B. Xue, B. Wang, B. Wu, B. Feng, C. Lu, C. Zhao, C. Deng, C. Zhang, C. Ruan, D. Dai, D. Chen, D. Ji, E. Li, F. Lin, F. Dai, F. Luo, G. Hao, G. Chen, G. Li, H. Zhang, H. Bao, H. Xu, H. Wa...

arXiv 2025
[7]

Dettmers, A

T. Dettmers, A. Pagnoni, A. Holtzman, and L. Zettlemoyer. Qlora: Efficient finetuning of quantized llms, 2023. URL https://arxiv.org/abs/2305.14314

arXiv 2023
[8]

llama.cpp

ggml.ai. llama.cpp. https://github.com/ggml-org/llama.cpp, 2023. GitHub repository

work page 2023

Show all 23 references

[9]

J. Hong, N. Lee, and J. Thorne. Orpo: Monolithic preference optimization without reference model, 2024. URL https://arxiv.org/abs/2403.07691

2024 arXiv
[10]

A. Q. Jiang, A. Sablayrolles, A. Mensch, C. Bamford, D. S. Chaplot, D. de las Casas, F. Bressand, G. Lengyel, G. Lample, L. Saulnier, L. R. Lavaud, M.-A. Lachaux, P. Stock, T. L. Scao, T. Lavril, T. Wang, T. Lacroix, and W. E. Sayed. Mistral 7b, 2023. URL https://arxiv.org/abs...

2023 arXiv
[11]

S. A. Laboratory. Opencompass. https://github.com/open-compass/opencompass, 2023 a . GitHub repository

2023
[12]

S. A. Laboratory. Xtuner. https://github.com/InternLM/xtuner, 2023 b . GitHub repository

2023
[13]

C. Liu, Z. Wang, S. Shen, J. Peng, X. Zhang, Z. Du, and Y. Wang. The chinese dataset distilled from deepseek-r1-671b. https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k, 2025

2025
[14]

V. B. Nguyen, J. Schlötterer, and C. Seifert. Ceval: A benchmark for evaluating counterfactual text generation, 2024. URL https://arxiv.org/abs/2404.17475

2024 arXiv
[15]

ollama. Ollama. https://github.com/ollama/ollama, 2023. GitHub repository

2023
[16]

Open webui

Open-WebUI. Open webui. https://github.com/open-webui/open-webui, 2023. GitHub repository

2023
[17]

Pipelines: Ui-agnostic openai api plugin framework

Open-WebUI. Pipelines: Ui-agnostic openai api plugin framework. https://github.com/open-webui/pipelines, 2024. GitHub repository

2024
[18]

Pku-saferlhf-single-dimension

PKU-Alignment. Pku-saferlhf-single-dimension. https://huggingface.co/datasets/PKU-Alignment/PKU-SafeRLHF-single-dimension, 2024. Huggingface repository

2024
[19]

Emollm-心理健康大模型

SmartFlowAI. Emollm-心理健康大模型. https://github.com/SmartFlowAI/EmoLLM, 2024. GitHub repository

2024
[20]

Touvron, T

H. Touvron, T. Lavril, G. Izacard, X. Martinet, M.-A. Lachaux, T. Lacroix, B. Rozière, N. Goyal, E. Hambro, F. Azhar, A. Rodriguez, A. Joulin, E. Grave, and G. Lample. Llama: Open and efficient foundation language models, 2023. URL https://arxiv.org/abs/2302.13971

2023 arXiv
[21]

Vaswani, N

A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin. Attention is all you need, 2023. URL https://arxiv.org/abs/1706.03762

2023 arXiv
[22]

H. Xie, Y. Chen, X. Xing, J. Lin, and X. Xu. Psydt: Using llms to construct the digital twin of psychological counselor with personalized counseling style for psychological counseling, 2024. URL https://arxiv.org/abs/2412.13660

2024 arXiv
[23]

Zhang, R

C. Zhang, R. Li, M. Tan, M. Yang, J. Zhu, D. Yang, J. Zhao, G. Ye, C. Li, and X. Hu. Cpsycoun: A report-based multi-turn dialogue reconstruction and evaluation framework for chinese psychological counseling, 2024. URL https://arxiv.org/abs/2405.16433

2024 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

J. Bai, S. Bai, Y. Chu, Z. Cui, K. Dang, X. Deng, Y. Fan, W. Ge, Y. Han, F. Huang, B. Hui, L. Ji, M. Li, J. Lin, R. Lin, D. Liu, G. Liu, C. Lu, K. Lu, J. Ma, R. Men, X. Ren, X. Ren, C. Tan, S. Tan, J. Tu, P. Wang, S. Wang, W. Wang, S. Wu, B. Xu, J. Xu, A. Yang, H. Yang, J. Yang, S. Yang, Y. Yao, B. Yu, H. Yuan, Z. Yuan, J. Zhang, X. Zhang, Y. Zhang, Z. Zh...

arXiv 2023

[2] [2]

Z. Cai, M. Cao, H. Chen, K. Chen, K. Chen, X. Chen, X. Chen, Z. Chen, Z. Chen, P. Chu, X. Dong, H. Duan, Q. Fan, Z. Fei, Y. Gao, J. Ge, C. Gu, Y. Gu, T. Gui, A. Guo, Q. Guo, C. He, Y. Hu, T. Huang, T. Jiang, P. Jiao, Z. Jin, Z. Lei, J. Li, J. Li, L. Li, S. Li, W. Li, Y. Li, H. Liu, J. Liu, J. Hong, K. Liu, K. Liu, X. Liu, C. Lv, H. Lv, K. Lv, L. Ma, R. Ma...

arXiv 2024

[3] [3]

H. Cao, Y. Wang, S. Jing, Z. Peng, Z. Bai, Z. Cao, M. Fang, F. Feng, B. Wang, J. Liu, T. Yang, J. Huo, Y. Gao, F. Meng, X. Yang, C. Deng, and J. Feng. Safedialbench: A fine-grained safety benchmark for large language models in multi-turn dialogues with diverse jailbreak attacks, 2025. URL https://arxiv.org/abs/2502.11090

arXiv 2025

[4] [4]

DeepPsy-Agent: A Stage-Aware and Deep-Thinking Emotional Support Agent System

K. Chen and Z. Sun. Deeppsy-agent: A stage-aware and deep-thinking emotional support agent system, 2025. URL https://arxiv.org/abs/2503.15876

work page Pith review arXiv 2025

[5] [5]

DeepSeek-AI, :, X. Bi, D. Chen, G. Chen, S. Chen, D. Dai, C. Deng, H. Ding, K. Dong, Q. Du, Z. Fu, H. Gao, K. Gao, W. Gao, R. Ge, K. Guan, D. Guo, J. Guo, G. Hao, Z. Hao, Y. He, W. Hu, P. Huang, E. Li, G. Li, J. Li, Y. Li, Y. K. Li, W. Liang, F. Lin, A. X. Liu, B. Liu, W. Liu, X. Liu, X. Liu, Y. Liu, H. Lu, S. Lu, F. Luo, S. Ma, X. Nie, T. Pei, Y. Piao, J...

arXiv 2024

[6] [6]

DeepSeek-AI, D. Guo, D. Yang, H. Zhang, J. Song, R. Zhang, R. Xu, Q. Zhu, S. Ma, P. Wang, X. Bi, X. Zhang, X. Yu, Y. Wu, Z. F. Wu, Z. Gou, Z. Shao, Z. Li, Z. Gao, A. Liu, B. Xue, B. Wang, B. Wu, B. Feng, C. Lu, C. Zhao, C. Deng, C. Zhang, C. Ruan, D. Dai, D. Chen, D. Ji, E. Li, F. Lin, F. Dai, F. Luo, G. Hao, G. Chen, G. Li, H. Zhang, H. Bao, H. Xu, H. Wa...

arXiv 2025

[7] [7]

Dettmers, A

T. Dettmers, A. Pagnoni, A. Holtzman, and L. Zettlemoyer. Qlora: Efficient finetuning of quantized llms, 2023. URL https://arxiv.org/abs/2305.14314

arXiv 2023

[8] [8]

llama.cpp

ggml.ai. llama.cpp. https://github.com/ggml-org/llama.cpp, 2023. GitHub repository

work page 2023