Zitcha

zitcha

Authors:
(1) Divyanshu Kumar, Enkrypt AI;
(2) Anurakt Kumar, Enkrypt AI;
(3) Sahil Agarwa, Enkrypt AI;
(4) Prashanth Harshangi, Enkrypt AI. Table of Links Abstract and 1 Introduction 2 Problem Formulation and Experiments 3 Experiment Set-up & Results 4 Conclusion and References A. Appendix 4 CONCLUSION Our work investigates the LLM’s safety against Jailbreak attempts. We have demonstrated how fine-tuned and quantized models are vulnerable to jailbreak attempts and stress the importance of using external guardrails to reduce this risk. Fine-tuning or quantizing model weights alters the risk profile of LLMs, potentially undermining the safety alignment established through RLHF. This could result from catastrophic forgetting, where LLMs lose memory of safety protocols, or the finetuning process shifting the model’s focus to new topics at the expense of existing safety measures. The lack of safety measures in these fine-tuned and quantized models is concerning, highlighting the need to incorporate safety protocols during the fine-tuning process. We propose using these tests as part of a CI/CD stress test before deploying the model. The effectiveness of guardrails in preventing jailbreaking highlights the importance of integrating them with safety practices in AI development. This approach not only enhances AI models but also establishes a new standard for responsible AI development. By ensuring that AI advancements prioritize innovation and safety, we promote ethical AI deployment, safeguarding against potential misuse and fostering a secure digital future. REFERENCES Zifan Wang Andy Zou. AdvBench Dataset, July 2023. URL https://github.com/ llm-attacks/llm-attacks/tree/main/data. Patrick Chao, Alexander Robey, Edgar Dobriban, Hamed Hassani, George J. Pappas, and Eric Wong. Jailbreaking Black Box Large Language Models in Twenty Queries. arXiv, October 2023. doi: 10.48550/arXiv.2310.08419. Edoardo Debenedetti, Giorgio Severi, Nicholas Carlini, Christopher A. Choquette-Choo, Matthew Jagielski, Milad Nasr, Eric Wallace, and Florian Tramer. Privacy Side Channels in Machine ` Learning Systems. arXiv, September 2023. doi: 10.48550/arXiv.2309.05610. Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. QLoRA: Efficient Finetuning of Quantized LLMs. arXiv, May 2023. doi: 10.48550/arXiv.2305.14314. Micah Gorsline, James Smith, and Cory Merkel. On the Adversarial Robustness of Quantized Neural Networks. arXiv, May 2021. doi: 10.1145/3453688.3461755. Kai Greshake, Sahar Abdelnabi, Shailesh Mishra, Christoph Endres, Thorsten Holz, and Mario Fritz. Not what you’ve signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection. arXiv, February 2023. doi: 10.48550/arXiv.2302.12173. Bing He, Mustaque Ahamad, and Srijan Kumar. PETGEN: Personalized Text Generation Attack on Deep Sequence Embedding-based Classification Models. In KDD ’21: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, pp. 575–584. Association for Computing Machinery, New York, NY, USA, August 2021. ISBN 978-1-45038332-5. doi: 10.1145/3447548.3467390. Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. LoRA: Low-Rank Adaptation of Large Language Models. arXiv, June 2021. doi: 10.48550/arXiv.2106.09685. Shuhei Kashiwamura, Ayaka Sakata, and Masaaki Imaizumi. Effect of Weight Quantization on Learning Models by Typical Case Analysis. arXiv, January 2024. doi: 10.48550/arXiv.2401. 17269. Siwon Kim, Sangdoo Yun, Hwaran Lee, Martin Gubri, Sungroh Yoon, and Seong Joon Oh. ProPILE: Probing Privacy Leakage in Large Language Models. arXiv, July 2023. doi: 10.48550/arXiv. 2307.01881. Aounon Kumar, Chirag Agarwal, Suraj Srinivas, Aaron Jiaxun Li, Soheil Feizi, and Himabindu Lakkaraju. Certifying LLM Safety against Adversarial Prompting. arXiv, September 2023. doi: 10.48550/arXiv.2309.02705. Thai Le, Suhang Wang, and Dongwon Lee. MALCOM: Generating Malicious Comments to Attack Neural Fake News Detection Models. IEEE Computer Society, November 2020. ISBN 978-1- 7281-8316-9. doi: 10.1109/ICDM50108.2020.00037. Yi Liu, Gelei Deng, Zhengzi Xu, Yuekang Li, Yaowen Zheng, Ying Zhang, Lida Zhao, Tianwei Zhang, and Yang Liu. Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study. arXiv, May 2023. doi: 10.48550/arXiv.2305.13860. Anay Mehrotra, Manolis Zampetakis, Paul Kassianik, Blaine Nelson, Hyrum Anderson, Yaron Singer, and Amin Karbasi. Tree of Attacks: Jailbreaking Black-Box LLMs Automatically. arXiv, December 2023. doi: 10.48550/arXiv.2312.02119. Long Ouyang, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul Christiano, Jan Leike, and Ryan Lowe. Training language models to follow instructions with human feedback. arXiv, March 2022. doi: 10.48550/arXiv.2203.02155. Xiangyu Qi, Yi Zeng, Tinghao Xie, Pin-Yu Chen, Ruoxi Jia, Prateek Mittal, and Peter Henderson. Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To! arXiv, October 2023. doi: 10.48550/arXiv.2310.03693. Traian Rebedea, Razvan Dinu, Makesh Narsimhan Sreedhar, Christopher Parisien, and Jonathan Cohen. NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails. ACL Anthology, pp. 431–445, December 2023. doi: 10.18653/v1/2023. emnlp-demo.40. Alexander Wei, Nika Haghtalab, and Jacob Steinhardt. Jailbroken: How Does LLM Safety Training Fail? arXiv, July 2023. doi: 10.48550/arXiv.2307.02483. Martin Weyssow, Xin Zhou, Kisub Kim, David Lo, and Houari Sahraoui. Exploring ParameterEfficient Fine-Tuning Techniques for Code Generation with Large Language Models. arXiv, August 2023. doi: 10.48550/arXiv.2308.10462. Yisong Xiao, Aishan Liu, Tianyuan Zhang, Haotong Qin, Jinyang Guo, and Xianglong Liu. RobustMQ: Benchmarking Robustness of Quantized Models. arXiv, August 2023. doi: 10.48550/ arXiv.2308.02350. Andy Zhou, Bo Li, and Haohan Wang. Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks. arXiv, January 2024. doi: 10.48550/arXiv.2401.17263. Sicheng Zhu, Ruiyi Zhang, Bang An, Gang Wu, Joe Barrow, Zichao Wang, Furong Huang, Ani Nenkova, and Tong Sun. AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models. arXiv, October 2023. doi: 10.48550/arXiv.2310.15140. Andy Zou, Zifan Wang, Nicholas Carlini, Milad Nasr, J. Zico Kolter, and Matt Fredrikson. Universal and Transferable Adversarial Attacks on Aligned Language Models. arXiv, July 2023. doi: 10.48550/arXiv.2307.15043. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. Authors: (1) Divyanshu Kumar, Enkrypt AI; (2) Anurakt Kumar, Enkrypt AI; (3) Sahil Agarwa, Enkrypt AI; (4) Prashanth Harshangi, Enkrypt AI. Authors: Authors: (1) Divyanshu Kumar, Enkrypt AI; (2) Anurakt Kumar, Enkrypt AI; (3) Sahil Agarwa, Enkrypt AI; (4) Prashanth Harshangi, Enkrypt AI. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 Problem Formulation and Experiments 2 Problem Formulation and Experiments 3 Experiment Set-up & Results 3 Experiment Set-up & Results 4 Conclusion and References 4 Conclusion and References A. Appendix A. Appendix 4 CONCLUSION Our work investigates the LLM’s safety against Jailbreak attempts. We have demonstrated how fine-tuned and quantized models are vulnerable to jailbreak attempts and stress the importance of using external guardrails to reduce this risk. Fine-tuning or quantizing model weights alters the risk profile of LLMs, potentially undermining the safety alignment established through RLHF. This could result from catastrophic forgetting, where LLMs lose memory of safety protocols, or the finetuning process shifting the model’s focus to new topics at the expense of existing safety measures. The lack of safety measures in these fine-tuned and quantized models is concerning, highlighting the need to incorporate safety protocols during the fine-tuning process. We propose using these tests as part of a CI/CD stress test before deploying the model. The effectiveness of guardrails in preventing jailbreaking highlights the importance of integrating them with safety practices in AI development. This approach not only enhances AI models but also establishes a new standard for responsible AI development. By ensuring that AI advancements prioritize innovation and safety, we promote ethical AI deployment, safeguarding against potential misuse and fostering a secure digital future. REFERENCES Zifan Wang Andy Zou. AdvBench Dataset, July 2023. URL https://github.com/ llm-attacks/llm-attacks/tree/main/data. Patrick Chao, Alexander Robey, Edgar Dobriban, Hamed Hassani, George J. Pappas, and Eric Wong. Jailbreaking Black Box Large Language Models in Twenty Queries. arXiv, October 2023. doi: 10.48550/arXiv.2310.08419. Edoardo Debenedetti, Giorgio Severi, Nicholas Carlini, Christopher A. Choquette-Choo, Matthew Jagielski, Milad Nasr, Eric Wallace, and Florian Tramer. Privacy Side Channels in Machine ` Learning Systems. arXiv, September 2023. doi: 10.48550/arXiv.2309.05610. Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. QLoRA: Efficient Finetuning of Quantized LLMs. arXiv, May 2023. doi: 10.48550/arXiv.2305.14314. Micah Gorsline, James Smith, and Cory Merkel. On the Adversarial Robustness of Quantized Neural Networks. arXiv, May 2021. doi: 10.1145/3453688.3461755. Kai Greshake, Sahar Abdelnabi, Shailesh Mishra, Christoph Endres, Thorsten Holz, and Mario Fritz. Not what you’ve signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection. arXiv, February 2023. doi: 10.48550/arXiv.2302.12173. Bing He, Mustaque Ahamad, and Srijan Kumar. PETGEN: Personalized Text Generation Attack on Deep Sequence Embedding-based Classification Models. In KDD ’21: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, pp. 575–584. Association for Computing Machinery, New York, NY, USA, August 2021. ISBN 978-1-45038332-5. doi: 10.1145/3447548.3467390. Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. LoRA: Low-Rank Adaptation of Large Language Models. arXiv, June 2021. doi: 10.48550/arXiv.2106.09685. Shuhei Kashiwamura, Ayaka Sakata, and Masaaki Imaizumi. Effect of Weight Quantization on Learning Models by Typical Case Analysis. arXiv, January 2024. doi: 10.48550/arXiv.2401. 17269. Siwon Kim, Sangdoo Yun, Hwaran Lee, Martin Gubri, Sungroh Yoon, and Seong Joon Oh. ProPILE: Probing Privacy Leakage in Large Language Models. arXiv, July 2023. doi: 10.48550/arXiv. 2307.01881. Aounon Kumar, Chirag Agarwal, Suraj Srinivas, Aaron Jiaxun Li, Soheil Feizi, and Himabindu Lakkaraju. Certifying LLM Safety against Adversarial Prompting. arXiv, September 2023. doi: 10.48550/arXiv.2309.02705. Thai Le, Suhang Wang, and Dongwon Lee. MALCOM: Generating Malicious Comments to Attack Neural Fake News Detection Models. IEEE Computer Society, November 2020. ISBN 978-1- 7281-8316-9. doi: 10.1109/ICDM50108.2020.00037. Yi Liu, Gelei Deng, Zhengzi Xu, Yuekang Li, Yaowen Zheng, Ying Zhang, Lida Zhao, Tianwei Zhang, and Yang Liu. Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study. arXiv, May 2023. doi: 10.48550/arXiv.2305.13860. Anay Mehrotra, Manolis Zampetakis, Paul Kassianik, Blaine Nelson, Hyrum Anderson, Yaron Singer, and Amin Karbasi. Tree of Attacks: Jailbreaking Black-Box LLMs Automatically. arXiv, December 2023. doi: 10.48550/arXiv.2312.02119. Long Ouyang, Jeff Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul Christiano, Jan Leike, and Ryan Lowe. Training language models to follow instructions with human feedback. arXiv, March 2022. doi: 10.48550/arXiv.2203.02155. Xiangyu Qi, Yi Zeng, Tinghao Xie, Pin-Yu Chen, Ruoxi Jia, Prateek Mittal, and Peter Henderson. Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To! arXiv, October 2023. doi: 10.48550/arXiv.2310.03693. Traian Rebedea, Razvan Dinu, Makesh Narsimhan Sreedhar, Christopher Parisien, and Jonathan Cohen. NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails. ACL Anthology, pp. 431–445, December 2023. doi: 10.18653/v1/2023. emnlp-demo.40. Alexander Wei, Nika Haghtalab, and Jacob Steinhardt. Jailbroken: How Does LLM Safety Training Fail? arXiv, July 2023. doi: 10.48550/arXiv.2307.02483. Martin Weyssow, Xin Zhou, Kisub Kim, David Lo, and Houari Sahraoui. Exploring ParameterEfficient Fine-Tuning Techniques for Code Generation with Large Language Models. arXiv, August 2023. doi: 10.48550/arXiv.2308.10462. Yisong Xiao, Aishan Liu, Tianyuan Zhang, Haotong Qin, Jinyang Guo, and Xianglong Liu. RobustMQ: Benchmarking Robustness of Quantized Models. arXiv, August 2023. doi: 10.48550/ arXiv.2308.02350. Andy Zhou, Bo Li, and Haohan Wang. Robust Prompt Optimization for Defending Language Models Against Jailbreaking Attacks. arXiv, January 2024. doi: 10.48550/arXiv.2401.17263. Sicheng Zhu, Ruiyi Zhang, Bang An, Gang Wu, Joe Barrow, Zichao Wang, Furong Huang, Ani Nenkova, and Tong Sun. AutoDAN: Interpretable Gradient-Based Adversarial Attacks on Large Language Models. arXiv, October 2023. doi: 10.48550/arXiv.2310.15140. Andy Zou, Zifan Wang, Nicholas Carlini, Milad Nasr, J. Zico Kolter, and Matt Fredrikson. Universal and Transferable Adversarial Attacks on Aligned Language Models. arXiv, July 2023. doi: 10.48550/arXiv.2307.15043. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. This paper is available on arxiv under CC BY-NC-SA 4.0 DEED license. available on arxiv

Increased LLM Vulnerabilities from Fine-tuning and Quantization: Conclusion and References

Large Language Models on Memory-Constrained Devices Using Flash Memory: Conclusion & Discussion

VEATIC: Familiarity and Enjoyment Ratings & References

VEATIC: পরিচিতি এবং উপভোগ রেটিং এবং রেফারেন্স

VEATIC: Vertrautheits- und Genussbewertungen und Referenzen

VEATIC: Calificaciones y referencias de familiaridad y disfrute

VEATIC : Notes et références de familiarité et de plaisir

VEATIC: परिचितता और आनंद रेटिंग और संदर्भ

VEATIC: 親しみやすさと楽しさの評価と参考文献

VEATIC: 친숙함과 즐거움 평가 및 참고자료

VEATIC: classificações e referências de familiaridade e prazer

VEATIC: рейтинги знакомства и удовольствия, а также ссылки

VEATIC: Aşinalık ve Eğlence Derecelendirmeler ve Referanslar

VEATIC: Đánh giá và tham khảo về sự quen thuộc và thích thú

VEATIC：熟悉度和享受度评级及参考

The Gods of Mars by Edgar Rice Burroughs, is part of the HackerNoon Books Series. You can jump to any chapter in this book here . DOOMED TO DIE DOOMED TO DIE For an instant I stood there before they fell upon me, but the first rush of them forced me back a step or two. My foot felt for the floor but found only empty space. I had backed into the pit which had received Issus. For a second I toppled there upon the brink. Then I too with the boy still tightly clutched in my arms pitched backward into the black abyss. We struck a polished chute, the opening above us closed as magically as it had opened, and we shot down, unharmed, into a dimly lighted apartment far below the arena. As I rose to my feet the first thing I saw was the malignant countenance of Issus glaring at me through the heavy bars of a grated door at one side of the chamber. “Rash mortal!” she shrilled. “You shall pay the awful penalty for your blasphemy in this secret cell. Here you shall lie alone and in darkness with the carcass of your accomplice festering in its rottenness by your side, until crazed by loneliness and hunger you feed upon the crawling maggots that were once a man.” That was all. In another instant she was gone, and the dim light which had filled the cell faded into Cimmerian blackness. “Pleasant old lady,” said a voice at my side. “Who speaks?” I asked. “’Tis I, your companion, who has had the honour this day of fighting shoulder to shoulder with the greatest warrior that ever wore metal upon Barsoom.” “I thank God that you are not dead,” I said. “I feared for that nasty cut upon your head.” “It but stunned me,” he replied. “A mere scratch.” “Maybe it were as well had it been final,” I said. “We seem to be in a pretty fix here with a splendid chance of dying of starvation and thirst.” “Where are we?” “Beneath the arena,” I replied. “We tumbled down the shaft that swallowed Issus as she was almost at our mercy.” He laughed a low laugh of pleasure and relief, and then reaching out through the inky blackness he sought my shoulder and pulled my ear close to his mouth. “Nothing could be better,” he whispered. “There are secrets within the secrets of Issus of which Issus herself does not dream.” “What do you mean?” “I laboured with the other slaves a year since in the remodelling of these subterranean galleries, and at that time we found below these an ancient system of corridors and chambers that had been sealed up for ages. The blacks in charge of the work explored them, taking several of us along to do whatever work there might be occasion for. I know the entire system perfectly. “There are miles of corridors honeycombing the ground beneath the gardens and the temple itself, and there is one passage that leads down to and connects with the lower regions that open on the water shaft that gives passage to Omean. “If we can reach the submarine undetected we may yet make the sea in which there are many islands where the blacks never go. There we may live for a time, and who knows what may transpire to aid us to escape?” He had spoken all in a low whisper, evidently fearing spying ears even here, and so I answered him in the same subdued tone. “Lead back to Shador, my friend,” I whispered. “Xodar, the black, is there. We were to attempt our escape together, so I cannot desert him.” “No,” said the boy, “one cannot desert a friend. It were better to be recaptured ourselves than that.” Then he commenced groping his way about the floor of the dark chamber searching for the trap that led to the corridors beneath. At length he summoned me by a low, “S-s-t,” and I crept toward the sound of his voice to find him kneeling on the brink of an opening in the floor. “There is a drop here of about ten feet,” he whispered. “Hang by your hands and you will alight safely on a level floor of soft sand.” Very quietly I lowered myself from the inky cell above into the inky pit below. So utterly dark was it that we could not see our hands at an inch from our noses. Never, I think, have I known such complete absence of light as existed in the pits of Issus. For an instant I hung in mid air. There is a strange sensation connected with an experience of that nature which is quite difficult to describe. When the feet tread empty air and the distance below is shrouded in darkness there is a feeling akin to panic at the thought of releasing the hold and taking the plunge into unknown depths. Although the boy had told me that it was but ten feet to the floor below I experienced the same thrills as though I were hanging above a bottomless pit. Then I released my hold and dropped—four feet to a soft cushion of sand. The boy followed me. “Raise me to your shoulders,” he said, “and I will replace the trap.” This done he took me by the hand, leading me very slowly, with much feeling about and frequent halts to assure himself that he did not stray into wrong passageways. Presently we commenced the descent of a very steep incline. “It will not be long,” he said, “before we shall have light. At the lower levels we meet the same stratum of phosphorescent rock that illuminates Omean.” Never shall I forget that trip through the pits of Issus. While it was devoid of important incidents yet it was filled for me with a strange charm of excitement and adventure which I think must have hinged principally on the unguessable antiquity of these long-forgotten corridors. The things which the Stygian darkness hid from my objective eye could not have been half so wonderful as the pictures which my imagination wrought as it conjured to life again the ancient peoples of this dying world and set them once more to the labours, the intrigues, the mysteries and the cruelties which they had practised to make their last stand against the swarming hordes of the dead sea bottoms that had driven them step by step to the uttermost pinnacle of the world where they were now intrenched behind an impenetrable barrier of superstition. In addition to the green men there had been three principal races upon Barsoom. The blacks, the whites, and a race of yellow men. As the waters of the planet dried and the seas receded, all other resources dwindled until life upon the planet became a constant battle for survival. The various races had made war upon one another for ages, and the three higher types had easily bested the green savages of the water places of the world, but now that the receding seas necessitated constant abandonment of their fortified cities and forced upon them a more or less nomadic life in which they became separated into smaller communities they soon fell prey to the fierce hordes of green men. The result was a partial amalgamation of the blacks, whites and yellows, the result of which is shown in the present splendid race of red men. I had always supposed that all traces of the original races had disappeared from the face of Mars, yet within the past four days I had found both whites and blacks in great multitudes. Could it be possible that in some far-off corner of the planet there still existed a remnant of the ancient race of yellow men? My reveries were broken in upon by a low exclamation from the boy. “At last, the lighted way,” he cried, and looking up I beheld at a long distance before us a dim radiance. As we advanced the light increased until presently we emerged into well-lighted passageways. From then on our progress was rapid until we came suddenly to the end of a corridor that let directly upon the ledge surrounding the pool of the submarine. The craft lay at her moorings with uncovered hatch. Raising his finger to his lips and then tapping his sword in a significant manner, the youth crept noiselessly toward the vessel. I was close at his heels. Silently we dropped to the deserted deck, and on hands and knees crawled toward the hatchway. A stealthy glance below revealed no guard in sight, and so with the quickness and the soundlessness of cats we dropped together into the main cabin of the submarine. Even here was no sign of life. Quickly we covered and secured the hatch. Then the boy stepped into the pilot house, touched a button and the boat sank amid swirling waters toward the bottom of the shaft. Even then there was no scurrying of feet as we had expected, and while the boy remained to direct the boat I slid from cabin to cabin in futile search for some member of the crew. The craft was entirely deserted. Such good fortune seemed almost unbelievable. When I returned to the pilot house to report the good news to my companion he handed me a paper. “This may explain the absence of the crew,” he said. It was a radio-aerial message to the commander of the submarine: “The slaves have risen. Come with what men you have and those that you can gather on the way. Too late to get aid from Omean. They are massacring all within the amphitheatre. Issus is threatened. Haste. “ZITHAD” “Zithad is Dator of the guards of Issus,” explained the youth. “We gave them a bad scare—one that they will not soon forget.” “Let us hope that it is but the beginning of the end of Issus,” I said. “Only our first ancestor knows,” he replied. We reached the submarine pool in Omean without incident. Here we debated the wisdom of sinking the craft before leaving her, but finally decided that it would add nothing to our chances for escape. There were plenty of blacks on Omean to thwart us were we apprehended; however many more might come from the temples and gardens of Issus would not in any way decrease our chances. We were now in a quandary as to how to pass the guards who patrolled the island about the pool. At last I hit upon a plan. “What is the name or title of the officer in charge of these guards?” I asked the boy. “A fellow named Torith was on duty when we entered this morning,” he replied. “Good. And what is the name of the commander of the submarine?” “Yersted.” I found a dispatch blank in the cabin and wrote the following order: “Dator Torith: Return these two slaves at once to Shador. “YERSTED” “That will be the simpler way to return,” I said, smiling, as I handed the forged order to the boy. “Come, we shall see now how well it works.” “But our swords!” he exclaimed. “What shall we say to explain them?” “Since we cannot explain them we shall have to leave them behind us,” I replied. “Is it not the extreme of rashness to thus put ourselves again, unarmed, in the power of the First Born?” “It is the only way,” I answered. “You may trust me to find a way out of the prison of Shador, and I think, once out, that we shall find no great difficulty in arming ourselves once more in a country which abounds so plentifully in armed men.” “As you say,” he replied with a smile and shrug. “I could not follow another leader who inspired greater confidence than you. Come, let us put your ruse to the test.” Boldly we emerged from the hatchway of the craft, leaving our swords behind us, and strode to the main exit which led to the sentry’s post and the office of the Dator of the guard. At sight of us the members of the guard sprang forward in surprise, and with levelled rifles halted us. I held out the message to one of them. He took it and seeing to whom it was addressed turned and handed it to Torith who was emerging from his office to learn the cause of the commotion. The black read the order, and for a moment eyed us with evident suspicion. “Where is Dator Yersted?” he asked, and my heart sank within me, as I cursed myself for a stupid fool in not having sunk the submarine to make good the lie that I must tell. “His orders were to return immediately to the temple landing,” I replied. Torith took a half step toward the entrance to the pool as though to corroborate my story. For that instant everything hung in the balance, for had he done so and found the empty submarine still lying at her wharf the whole weak fabric of my concoction would have tumbled about our heads; but evidently he decided the message must be genuine, nor indeed was there any good reason to doubt it since it would scarce have seemed credible to him that two slaves would voluntarily have given themselves into custody in any such manner as this. It was the very boldness of the plan which rendered it successful. “Were you connected with the rising of the slaves?” asked Torith. “We have just had meagre reports of some such event.” “All were involved,” I replied. “But it amounted to little. The guards quickly overcame and killed the majority of us.” He seemed satisfied with this reply. “Take them to Shador,” he ordered, turning to one of his subordinates. We entered a small boat lying beside the island, and in a few minutes were disembarking upon Shador. Here we were returned to our respective cells; I with Xodar, the boy by himself; and behind locked doors we were again prisoners of the First Born. About HackerNoon Book Series: We bring you the most important technical, scientific, and insightful public domain books. This book is part of the public domain. Edgar Rice Burroughs (1993). The Gods of Mars. Urbana, Illinois: Project Gutenberg. Retrieved October 2022 https://www.gutenberg.org/cache/epub/64/pg64-images.html This eBook is for the use of anyone anywhere at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this eBook or online at www.gutenberg.org , located at https://www.gutenberg.org/policy/license.html .

DOOMED TO DIE

Blockchain Vs. BITCOIN. This Paid Story is brought to you by . INGOT Coin Lets get one thing clear; Blockchain technology is here to stay. Whether BTC, ETH or any other Cryptocurrency will in price is irrelevant to the bigger picture. Just look at the and the opportunities it has created, and the opportunities it is still creating. Blockchain technology can crash only if the internet crashes, and do you really think we would ever give up the internet? How can we capitalize on that fact? Easy, collaboration and a increase or decrease Internet more unified approach to applying and using Blockchain. You probably stumble upon many opinions stating that Bitcoin price will reach new highs from 30,000 to extreme prices of 100,000 and more. Everyone is waiting and expecting the price to soar this year but why is that? Mining supply is diminishing; which will increase demand? Growth in trading Bitcoins and other altcoins which is increasing mediums of exchange? Maybe.. but could it be that people are believing and putting their faith in Blockchain? Competition is Killing Blockchain I mean c’mon you can see it right? Whether you are an you can see this inefficiency. Everyone is bored with the same strategy and the same result from the ICO trend. Everyone is looking at it the wrong way; there is too much potential that is being lost because everyone is trying to finish first. Can you truly say that any ICO was able to use Blockchain to its fullest even after collecting millions in funds? investor, company, ICO or even short-term trader, The competition in the market is killing many great ideas that would enhance and improve industries before they even begin. There can be hundreds of similar projects in different regions, but they are in no way identical. Projects are being judged based on different ratings, PR, team size and weaknesses rather than strengths, which makes investors and participants look for the next similar ICO. No matter how good an ICO is, or how elaborate the team is, they can only go so far; now put that same ICO with multiple other ones and have them cooperate, and only then will there be no limits or restrictions. The INGOT Coin Group; Creating Real Utility Value We practice what we preach and if we believe in collaboration and the greater good of it, then collaboration is what we will be. INGOT Coin intends to bridge the Crypto markets to the traditional markets but not only through the ICO and our IC Ecosystem. INGOT Coin has taken an initiative to bring on Board different ICOs with great ideas and teams that strive to provide innovative services to their communities; creating the INGOT Coin Group. share a similar vision. We are here to explore multiple ways to further enhance all our communities. By combining the different Utility functions of our tokens, as well as exploring methods of integration between the different ecosystems will create a new meaning to Utility. Instead of having a single or a couple of utility benefits from one ICO, why not have many more benefits from unified and different ICOs ? INGOT Coin Group and Partners As of now this does not exist in the Crypto markets, financial markets or you know what, in any market. Joining one ICO will provide you with innovative services but even at this point the services are limited. That is why people also invest in multiple other ICOs and gain from different services as well, however the truth of the matter is even when that is done there really is no bridge or implementation between the ICOs. Meet Our Partners After speaking to various ICOs and Organizations about our collaborative vision, we noticed that we were all aligned in forming an alliance and gathering all our resources to innovate and provide further for the community! This vision and initiative started as soon as I connected and began formally cooperating realizing the great synergy that exists between the two companies and the wide room available for further innovation and development. NGOT Coin and Trade.io will be advised by Trade.io to enhance their ICO and leverage their established stakeholders with T expert consulting services to maximize value to INGOT’s ICO participants. They both share similar attributes in that they work with respectable universities, quality ICOs and respectable individuals within the Blockchain sector. More information on how the two companies will cooperate will be released soon. INGOT Coin is pleased to announce an advisory partnership with Trade.io . INGOT Coin rade.io’s is pleased to announce a partnership with The partnership ensures liquidity for the Modern Finance Chain payment solution by delivering access to various liquidity pools through INGOT Coin and its currency partners. Additionally, MF Chain is able to provide retail processing integration for INGOT Coin. INGOT Coin and MF Chain share a common vision of uniting and providing users with a multitude of solutions in an combined effort to drive Cryptocurrency adoption. Modern Finance Chain (MFCHAIN) : Modern Finance Chain INGOT Coin . INGOT Coin is pleased to announce a partnership with ZICHAIN . More details on the partnership will be announced early next week. INGOT Coin is pleased to announce a partnership with Stella . More details on the partnership will be announced early next week. INGOT Coin is pleased to announce a partnership with HalalChain . More details on the partnership will be announced early next week. INGOT Coin is pleased to announce a possible partnership with black.insure . More details on the partnership will be announced early next week. INGOT Coin is pleased to announce a partnership with Bineuro . More details on the partnership will be announced early next week. There are others who will be announced soon as well. We are ready to innovate and make history.

What’s better than Bitcoin at $100,000?

Open source hosting platforms like GitHub, GitLab, BitBucket and SourceForge are becoming ever more popular for developers to host and review code, to manage projects and build software. With a user base of over 28 million developers and more than 1.8 million businesses and organizations using its platform, is considered the most popular project in its category. Just recently it acquired GitHub. Some say the takeover is a good thing, because now GitHub can now grow into an even bigger project and connect or collaborate with other existing Microsoft brands, while others, especially GitHub developers have trust issues with Microsoft for being a long-term opponent to open source software. Added to the fact that these kind of takeovers enable tech-giants to generate even more power to control the industry, this doesn’t sit well with a lot of GitHub users. GitHub Microsoft has announced Benefits for developers using open source hosting platforms are the ability to showcase their work, documentation, ability to contribute to other (open source) projects, managing options and integration options. On the downside these open source platforms are all centralized businesses and they do not always seem to have their customers best interest at heart. They are after all intended to make money on their business model. Developers can use the platform but their intellectual property rights as well as the lack of being rewarded for their input are downsides to looking out for the users that build the platform. aims to change that, by developing and launching a decentralized, autonomous community powered by Blockchain. HitChain ☑️ HitChain proposes to build a decentralized autonomous organization of blockchain community developers and improve upon the existing open source ecosystem that is, in their view, ineffective, unfair and centralized. HitChain intends to create the worlds largest software development and self-regulated, collaboration program through the blockchain. Features that HitChain wants to implement into their ecosystem to improve on existing open source hosting projects like GitHub, GitKraken and GitLab are: ➡️ HitChain’s main goal is to accept all developers wanting to interact in a decentralized infrastructure project, while formalizing fair profit distribution to realize investor’s intrinsic value. ➡️ Centralized projects are vulnerable to security breaches, technical failures and they can have scalability issues. HitChain can solve these issues with their decentralized autonomous organization. ➡️ By the decentralized nature of the HitChain project, developers work can be safely stored through the implementation of smart contracts. This way the developers intellectual property rights are verified and safeguarded through the blockchain. ➡️ In the centralized open source hosting ecosystems, developers aren’t rewarded for their work they contribute to the platform. HitChain will change this by installing the digital currency protocol. This protocol will incentivize developers based on the designated algorithm. By implementing this feature, HitChain ensures fairness, accountability, autonomy and justice by enabling developers to transact and earn in an algorithmic market freely. ➡️ A trustworthy, transparent competency, appraisal and credit system will be established to collect and manage developer’s data. This appraisal system will be cost-efficient. ➡️ Because of the implementation of blockchain technology, transactions executed on the HitChain network are transparent and publicly traceable. ➡️ The open source platform will be maintained and updated by the developers in the HitChain community, and all major decisions will be based on the community voting system that is supervised by an independent committee. Ever since its source code went public on GitHub, the HitChain project has already had 15 different contributors upload 718 commits over the course of a few weeks. Until now 349 developer stars has been given. Stars are a way for GitHub users to bookmark a repository for future reference. ☑️ According to ICO review website , HitChain raised a total amount of 2,300,000 USD in their token sale that launched or ended at the 24th of July. A total token supply of 102,400,000,000 HIT tokens have been minted of which 30% were sold in the private and public tokensale, 40% will go to future mining rewards, 5% will go towards their marketing budget, while the other 25% is divided between the team and the foundation (15% vs. 10%). HIT tokens are initially launched , using the Ethereum blockchain. After HitChains main net launch ERC-20 HIT tokens can be exchanged for native HIT tokens. ICOdrops as ERC-20 tokens HitChain has had its global exchange listing debut on digital asset exchange , where you can now trade HitChain against the following trading pairs HIT/ BTC and HIT/ ETH. Huobi Global ☑️ Team, advisers and investorsThe HitChain team is comprised of industry professionals, mainly from the Asian region, that have extensive experience in the field of open source products, and have worked at some of the worlds’ well-known companies, like Microsoft, RedHat, Openstack and OSChina. Two of China’s top blockchain experts, Jia Tian, Chief Scientist at Cortex and Haobo Ma, fouder of AELF Blockchain, are HitChain project advisers. The list of institutional investors on this project is pretty impressive. I will highlight three of them:➡️ ASIC-miner manufacturer and supplier has become an established business in the cryptocurrency industry.➡️ Huobi Capital is the investment arm of one of the worlds’ leading global digital asset financial service providers, Huobi Group.➡️ Venture Capital firm invests in companies based on their innovative technologies and business models with market appeal. BitMain Draper Dragon ☑️ As I see it, HitChain, which is still in its early stage of development, has a lot of upside potential, especially when they focus more on getting recognition from the western markets. Some points of improvements are the bilingual Telegram channel. A Chinese/ English chat is confusing, I would suggest splitting it into two separate channels. I would have liked to see the HitChain team members profiles on LinkedIn, but unfortunately they were not available. HitChain seems to have focused primarily on gaining a market share in the Asian market as there is not a lot of public information available in English. I would like to see some more project information in the form of press releases, Twitter posts and Medium articles. The recent takeover of GitHub by Microsoft has some developers looking for another go-to open source platform and being the new kid on the block with the added benefits like a reward system and intellectual property registration through the blockchain, HitChain can profit from that. Only time will tell whether or not HitChain will succeed in surpassing GitHub as being the worlds’ largest open source hosting platform. If you would like to know more about the project you can visit their , read their or join their or to learn about the latest updates. website whitepaper twitter Telegram channel ☑️ As of the 3rd of August, HitChain is trading at Huobi Global, with the trading pairs HIT/BTC and HIT/ETH being available. Huobi Global is one of the worlds’ leading digital asset exchanges. If you don’t have a trading account on Huobi Global yet, you can sign up, using link. By signing up you will automatically get approved for trading on (HADAX). this Huobi Autonomous Digital Asset Exchange Full disclosure : I do not own any HitChain tokens. This article is not intended as investment advice. It is just my personal opinion about the HitChain project. You should always do your own research. Huobi Pro rewards me for writing this article and supports me for ventilating my own personal opinion. Subscribe to my channels , and if you like my articles and would like to be informed about Blockchain, cryptocurrency projects and news. You can also read my articles on . If you have any questions about this article, please comment in the comment section below. Thank you!  LindaCrypto Steemit Medium Twitter LinkedIn

HitChain: the Community Powered Blockchain and newest edition to Huobi Global

Loss of Amazon Rainforest Resilience: Author Contributions

ROI4Presenter is excited to announce its transformation into in April 2024. Pitch Avatar Pitch Avatar is an AI-powered platform that makes your slides interactive and helps your content achieve its goals: grow leads, demo calls, and user engagement. Increase sales, marketing, onboarding, and training effectiveness - that's the mission of the personalized AI assistant, which saves you time and works 24/7 on your behalf worldwide in multiple languages. A bit of facts In 2023, Pitch Avatar became the 3rd product of the day on Product Hunt with its AI assistant in slides' delivery. After that, the product gained international recognition, presenting at five international summits, including the Web Summit in Lisbon. In just one year, it attracted 5000+ users, added 150+ new features, and saw website traffic soar to 60K. Users of Pitch Avatar especially value the ability of the product to translate presentations on the go, add voice-overs, or create AI avatars that present on their behalf. They also value the possibility of tracking the effectiveness & the goals achieved, as well as getting notifications when someone joins the session and the possibility of joining the session online. The platform is the most frequently used by: Sales professionals: Enhance outreach efforts and get more demo calls. Marketers: to generate more leads. HR managers: for initial candidates interviewing. Coaches and lecturers: to distribute their educational materials. Startups: to make their pitches outstanding. Would you like to try it? It's easy! Starting April 2024, Pitch Avatar has also updated its pricing policy and offered a 7-day free trial to experience all the features. The Trial includes unlimited AI-generated text scripts and voice-overs, 3 minutes for avatar generation, the ability to upload up to 10 presentations, the option to generate and share up to 50 links, and access in-depth analytics. If you need more time to try, you can and get access to an extended Trial. Book a demo “Our rebranding process is more than just a transformation in our name and identity. It's about enhancing the quality of our service. With Pitch Avatar, you can expect to access even more useful and highly effective features. We believe that Pitch Avatar will redefine the way presentations are made and delivered” - says Victoria Abed, CRO, Pitch Avatar. Imagine a future where your AI assistant helps you with routine tasks and accelerates goal achievement, allowing you to focus on more important things. With Pitch Avatar, that future is already here. Start your. Start your ! free Trial today Contacts us: info@pitchavatar.com support@pitchavatar.com This story was distributed as a press release under HackerNoon’s Business Blogging Program. Learn more about the program here.

ROI4Presenter Becomes Pitch Avatar: From Online Presentations to an AI-Based Platform

ROI4 উপস্থাপক পিচ অবতার হয়ে উঠেছে: অনলাইন উপস্থাপনা থেকে একটি AI-ভিত্তিক প্ল্যাটফর্মে

ROI4Presenter wird zum Pitch-Avatar: Von Online-Präsentationen zu einer KI-basierten Plattform

ROI4Presenter se convierte en el avatar de presentación: de presentaciones en línea a una plataforma basada en inteligencia artificial

ROI4Presenter devient Pitch Avatar : des présentations en ligne à une plate-forme basée sur l'IA

ROI4Presenter बना पिच अवतार: ऑनलाइन प्रेजेंटेशन से लेकर AI-आधारित प्लेटफॉर्म तक

ROI4Presenter がピッチアバターに: オンラインプレゼンテーションから AI ベースのプラットフォームへ

ROI4Presenter가 프레젠테이션 아바타가 되었습니다: 온라인 프레젠테이션에서 AI 기반 플랫폼까지

ROI4Presenter se torna o avatar do pitch: de apresentações on-line a uma plataforma baseada em IA

ROI4Presenter становится питч-аватаром: от онлайн-презентаций к платформе на базе искусственного интеллекта

ROI4Presenter Satış Konuşması Avatarına Dönüşüyor: Çevrimiçi Sunumlardan Yapay Zeka Tabanlı Platforma

ROI4Presenter trở thành Avatar quảng cáo chiêu hàng: Từ bài thuyết trình trực tuyến đến nền tảng dựa trên AI

ROI4Presenter 成为 Pitch Avatar：从在线演示到基于人工智能的平台

CAMBRIDGE 1828-1831

The Potwell Inn

Stories, News & Mentions about ZITCHA

Zitcha’s Stories on HackerNoon