DETAILED ACTION
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 03/09/2026 has been entered.
The application contains claims 1-20 all examined and rejected.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
It is acknowledged that claims 1, 4, 11, 14 and 20 were amended.
Response to Arguments
Applicant's arguments filed on 03/09/2026 regarding claims 1-20 being rejected under 35 USC 101 for being directed to an abstract idea without significantly more have been fully considered but they are not persuasive. Regarding the claims being evaluated under Prong one (Step 2A): “Applicant respectfully submits that the recited claim language cannot fairly be characterized as a mental process because the claim requires a computer implemented process that operates a large language model using modified prompts that are formed from retrieved context passages, with a measured and progressively increased context quantity. Applicant's claims are not limited to a person evaluating information and making a judgment. The claim requires the use of a large language model that generates an output for the query and generates a subsequent output for the modified prompt. The claims also recite forming modified prompts that include a measured number of context passages. The claim further requires iterative re-querying of the large language model using subsequent modified prompts in which the count is increased until an acceptable output is achieved.” Examiner respectfully submits that the claims recite the use of a large language model and a computer. Which, are recitation of generic computer components and under Prong one (Step 2A), and covers performance of the highlighted limitation(s) in the mind but for the recitation of generic computer components. That is, other than reciting “device,” “apparatus”, “network”, “a memory,” “a processor,” “non-transitory, computer-readable medium storing program instruction”, nothing in the claim element(s) precludes the step(s) from practically being performed in the human mind using observation, evaluation, judgment, and opinion. As such, the claim(s) falls within the “Mental Processes,” grouping of abstract ideas. Specifically, the core of the claim is retrieving relevant passages, evaluate whether the LLM’s answer is unacceptable, modify the prompt to include some the retrieved passages, and iteratively increate the amount of context until an acceptable answer is observed. While the steps describe retrieving context passages and providing the prompt using computer components and the use of an LLM, the limitations are merely instructions to implement an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea - see MPEP 2106.05(f). And fetching data, providing data input and receiving data output (i.e. mere data gathering, input and output recited at a high level of generality and thus are insignificant extra-solution activity - see MPEP 2106.05(g).
As such, under Step 2A Prong 2: The judicial exception(s) are not integrated into a practical application. And under Step 2B: The claim(s) does not include additional element(s) that are sufficient to amount to significantly more than the judicial exception.
Regarding Applicant’s argument related to technical improvement, “The claim also provides a technological improvement in the operation of systems that use retrieval augmented generation with large language models. Retrieval augmented generation has processing costs that grow with the amount of context provided to the model. Applicant's independent claims control and meter that cost by providing a measured amount of retrieved context and increasing that amount only as needed to reach an acceptable output. The claims therefore improve how the large language model is operated by controlling the context budget through an iterative mechanism defined by the measured parameter of the count. The claim also improves data handling and data exposure control within the prompting pipeline by limiting the amount of retrieved knowledge base content inserted into each modified prompt and expanding that content only when needed. This controlled expansion reduces unnecessary processing of retrieved passages and reduces unnecessary inclusion of retrieved repository content in model inputs during unsuccessful iterations.” Examiner suggest including limitations from claim 4 or 14 into the independent claim(s). The limitation provides the solution that enables the improved data handling and data exposure, reduces unnecessary processing of retrieved passages and reduces unnecessary inclusion of retrieved repository content in model inputs.
Applicant’s arguments with respect to the amended claims 1, 11 and 20 in view of Wang have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Allowable Subject Matter
Claim(s) 4 and 14 objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claim(s) 1-3, 5-13, 15-20 is/are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Claims 1-10, are method claims. Claims 11-19, are system claims. Claim 20 is a non-transitory computer readable medium claim. Therefore, claims 1-20 are directed to either a process, machine, manufacture, or composition of matter.
Step 2A Prong 1:
Claim 1, 11 and 20 recites the following limitation(s):
fetching
determiningjudgement which can be reasonably performed in one’s mind or with the aid of pencil and paper)
forming,
and increasing,
Claims 2, 3, 5-10, 12, 13, 15-19 are dependent on claims 1 and 11 respectively, and are evaluated similarly to claims 1, 11 and 20 above, which further are directed to a Mental process of evaluation and judgement which can be reasonably performed in one’s mind or with the aid of pencil and paper)
Accordingly, under its broadest reasonable interpretation, covers performance of the highlighted limitation(s) in the mind but for the recitation of generic computer components. That is, other than reciting “device,” “apparatus”, “network”, “a memory,” “a processor,” “non-transitory, computer-readable medium storing program instruction”, nothing in the claim element(s) precludes the step(s) from practically being performed in the human mind using observation, evaluation, judgment, and opinion. As such, the claim(s) falls within the “Mental Processes,” grouping of abstract ideas. Therefore, the claim(s) recites an abstract idea.
Step 2A Prong 2: The judicial exception(s) are not integrated into a practical application. The claim(s) recites the following additional elements:
Claim 1, recites: by a device; Claim 11, recites: apparatus, network interfaces, network, processor, memory; Claim 20 recites: non-transitory, computer-readable medium (all of which are recited at high level of generality and amounts to no more than mere instructions to apply the exception using a generic computer)
Claims 1, 12 and 20 further recites:
fetching, by a device…; providing, by the device… (which is equivalent to mere data gathering and output recited at a high level of generality, and thus are insignificant extra-solution activity)
providing, by the device the modified prompt as an input to the large language model to generate a subsequent output, (which is equivalent to mere data gathering and input recited at a high level of generality, and thus are insignificant extra-solution activity)
regarding the use of “large language model” recited in the claims, the large language model is used to generally apply the abstract idea without limiting how the large language model functions. The large language model is described at a high level such that it amounts to using a computer with a generic large language model to apply the abstract idea.
Claims 2, 3, 5-10, 12, 13, 15-19 are dependent on claims 1 and 11 respectively, and are evaluated similarly to claims 1, 11 and 20 above.
Accordingly, these additional element does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea. Therefore, the claim(s) are directed to an abstract idea.
Step 2B: The claim(s) does not include additional element(s) that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element(s) amounts to no more than mere instructions to apply the exception using a generic computer and thus are mere instructions to apply an exception using a generic computer component-see MPEP 2106.05(f). Also, the additional element(s) amounts to no more than mere data gathering and output recited at a high level of generality and thus are insignificant extra-solution activity - see MPEP 2106.05(g). Therefore, the additional element(s) are not indicative of an inventive concept (aka “significantly more”). The claim(s) are not patent eligible.
Examiner recommends incorporating claim(s) 4 or 14 into the independent claims to overcome the 101 rejections.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1, 5-11, 15-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang et al. (US 20230368268 A1) in view of Akari et al. “SELF-RAG: LEARNING TO RETRIEVE, GENERATE, AND CRITIQUE THROUGH SELF-REFLECTION” arXiv:2310.11511v1 [cs.CL] 17 Oct 2023.
Regarding claim 1, Wang discloses:
a method, comprising: fetching, by a device, relevant context responsive to a particular query initially submitted into a large language model without additional context, wherein the relevant context , at least by (paragraph [0061-0062, 0070] describes a user query as the query initially submitted into multi-model recommender system (e.g. a particular query initially submitted into a large language model without additional context), paragraph [0112] describe fetching relevant context responsive to a particular query, as recommended items 808 820 (see also Fig. 8A and 8B), paragraph [0063-0064] further describes the recommended items being from a product catalog (e.g. a knowledge based repository), paragraph [0115] further describes how the recommender system received the selected recommended items in 820 and augments the initial user query to provide an updated recommendation)
determining, by the device, that an output from the large language model for the particular query is an unacceptable output, at least by (paragraph [0082, 0113-0114] Fig. 8A Ref. 816, describes determining that the output Ref. 806 and 808 is an unacceptable output, as no items were selected and user response indicating a different context from the output and that the user wants to receive more recommendations)
forming, by the device at least by (paragraph [0114-0115] and Fig. 8B, two selections are made out of the three recommended items and are provided back to the recommender system to augment the initial query)
providing, by the device modified prompt as an input to the large language model to generate a a subsequent output, at least by (paragraph [0114-0115] and Fig. 8B, two selections are made out of the three recommended items and are provided back to the recommender system to augment the initial query (e.g. providing a select subset of the relevant context along with the particular query as an input to the large language model) and provide a subsequent output shown in Fig. 8C Ref. 824 and 826, which was in response to the unacceptable recommendations indicated by “user query 816 indicates that the user wants to receive more recommendations. As shown, the user did not select any of the items 808 as part of continuing to search for more recommendations,” (paragraph [0113).)
and increasing, by the device, progressively and responsive to subsequent unacceptable outputs from the large language model, an amount of relevant context in each subsequent subset iteratively provided to the large language model until an acceptable output is achieved, wherein increasing comprises increasing the count of context of passages included in subsequent modified prompts to include additional ones of the plurality of contest passages, at least by (paragraph [0100, 0114-0115] describes repeating the process of “increasing the amount of context” each time the user request additional items to recommend, the progressive increased context is shown by the recommendations going from women pants to men’s pants, to recommendations related to selected pants and Khaki colors and fit, and acceptable output is indicated when the user “add the item to an online shopping cart”)
But Wang fails to describe: the relevant context comprising a plurality of context passages; …comprising a count of context passages included in the modified prompt, the context passages being selected from the plurality of context passages sorted by relevance to the particular query; wherein increasing comprises increasing the count of context of passages included in subsequent modified prompts to include additional ones of the plurality of contest passages.
However, Akari teaches the above limitations at least by (Sec. 3.2.2, “For each segment yt ∈ y, we run C to assess whether additional passages could help to enhance generation. If retrieval is required, the retrieval special token Retrieve =Yes is added, and R retrieves the top K passages, D” where, “additional passages could help to enhance generation. If retrieval is required” describes recited limitation “responsive to the unacceptable output”, segment/passages describes recited limitation “plurality of context passages”, where in yt, t is the count of each segment/additional passage, and top K passages describes “selected from the plurality of context passages sorted by relevance to the particular query”. Further “For each passage, C further evaluates whether the passage is relevant and predicts ISREL . If a passage is relevant, C further evaluates whether the passage supports the model generation and predicts ISSUP . Critique tokens ISREL and ISSUP are appended after the retrieved passage or generations. At the end of the output, y (or yT), C predicts the overall utility token ISUSE , and an augmented output with reflection tokens and the original input pair is added to Dgen,” describes recited “increasing comprises increasing the count of context of passages included in subsequent modified prompts to include additional ones of the plurality of contest passages.”
Therefore, before the effective filing date of the invention it would have been obvious to one of ordinary skill in the art to combine the system of Wang with Akari to introduce[s] iterative prompt engineering to improve the factuality of LLM generations (Akari, Sec 4.2)
As per claim 5, claim 1 is incorporated and Wang further describes:
wherein the particular query comprises a user query, at least by (paragraph [0060] “the user interface 502 may also include a text input field, via which a user may input a query.”)
As per claim 6, claim 1 is incorporated and Wang further describes:
wherein determining the unacceptable output comprises: determining user dissatisfaction with the output, at least by (paragraph [0082, 0113-0114] Fig. 8A Ref. 816, describes determining that the output Ref. 806 and 808 is an unacceptable output, as no items were selected and user response indicating a different context from the output and that the user wants to receive more recommendations)
As per claim 7, claim 6 is incorporated and Wang further describes:
wherein determining user dissatisfaction comprises: receiving a user selection regarding satisfaction, at least by (paragraph [0114] Fig. 8B Ref. 820, describes the user selects two items of the items 820 (as illustrated by the ‘X’s in the upper-right corner of two of the recommended items 820) and inputs a user query 822. The selection of the two items indicates satisfaction, paragraph [0115] Fig. 8C Ref. 828, adding selected item to cart also indicates satisfaction9)
As per claim 8, claim 6 is incorporated and Wang further describes:
wherein determining user dissatisfaction comprises: receiving a conversational indication from a user regarding satisfaction, at least by (paragraph [0114] Fig. 8B Ref. 822, describes the user selects two items of the items 820 (as illustrated by the ‘X’s in the upper-right corner of two of the recommended items 820) and inputs a user query 822 the user query “These look good…but…” is a conversational indication from a user regarding satisfaction
As per claim 9, claim 1 is incorporated and Wang further describes:
determining the unacceptable output comprises: receiving an indication from a verification system that the output has been deemed unusable, at least by (paragraph [0082, 0113-0114] Fig. 8A Ref. 816, describes determining that the output Ref. 806 and 808 is an unacceptable output, as no items were selected and user response indicating a different context from the output and that the user wants to receive more recommendations)
As per claim 10, claim 9 is incorporated and Wang further describes:
wherein the verification system is external to the device, at least by (paragraph [0082, 0113-0114] Fig. 8A Ref. 816, describes determining that the output Ref. 806 and 808 is an unacceptable output, as no items were selected and user response indicating a different context from the output and that the user wants to receive more recommendations, where the user and user device (ref. 514) is external to the recommendation system (ref. 104) (see. Fig. 5))
Claims 11, 14-19 recite equivalent claim limitations as claims 1, 4-9 above, except that they set forth the claimed invention as a system; Claim 20 recite equivalent claim limitations as claim 1 above, except that they set forth the claimed invention as a non-transitory computer readable medium, as such they are rejected for the same reasons as applied hereinabove.
Claim(s) 2 and 12 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang and Akari in view of Fabian et al. (US 20240303440 A1).
As per claim 2, claim 1 is incorporate and Wang further describes:
further comprising: sorting a plurality of subsets of the relevant context according to relevance, at least by (paragraph [0036] which describes recommended products as a plurality of portions of the relevant context, and describe ranking the products, further in paragraph [0119-0120] ranking (e.g. sorting) the plurality of search results/recommended product)
But Wang fails to specifically describes: providing the select subset and each subsequent subset iteratively in order of relevance.
However, Fabian discloses the above limitation at least by (paragraph [0030] “determines an order in which to display the suggestions based… reply may suggest three formulas and indicate that one of the formulas is the most relevant” paragraph [0037] “application may order the alternative suggestions… present the alternative suggestions in the task pane”; See para. 0109-0112, Fig. 10A-10C, describes how each select portion is being provided iteratively based on relevant order as the user input are iteratively received)
Therefore, before the effective filing date of the invention it would have been obvious to one of ordinary skill in the art to combine Wang with the ability for the system to provide ordered suggestions to prompt the large language model as disclosed by Fabian to “to improve the performance of the LLM without overwhelming the LLM… by balance[ing] prompt size …with providing sufficient information to generate a useful response (Fabian, 0042).
Claim 12 recite equivalent claim limitations as claim 2 above, except that they set forth the claimed invention as a system as such they are rejected for the same reasons as applied hereinabove.
Claim(s) 3 and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Wang and Akari in view of Mehrotra et al. (US 20250148477 A1).
As per claim 3, claim 1 is incorporate but Wang fails to describes:
further comprising: tracking a final amount of context used to achieve the acceptable output; and using the final amount of context in response to a subsequent query topically similar to the particular query.
However, Mehrotra discloses the above limitation at least by (paragraph [0053] “the context retrieval component 206 can identify any previous chat histories 312 stored in the archived chat histories 312 that were directed to a customer support issue determined to be similar to the issue described by the prompt 306 currently being processed, and add these similar chat histories to the prompt 306 as chat history data 308. These similar chat histories can include information regarding how technical support issues similar to that described in the prompt 306 were resolved in the past, as well as metrics regarding how well the resolutions proposed by the model 226 satisfied the users' issues (e.g., in the form of user feedback or ratings)” and paragraph [0056] further describes how such tracked prompts leading to satisfying users’ issues are used in the model responses to similar prompts in the future (e.g. response to a subsequent query topically similar to the particular query))
Therefore, before the effective filing date of the invention it would have been obvious to one of ordinary skill in the art to combine Wang with the ability for the system to track the performance of responses to user prompts provided by Mehrotra to “refine or enhance the initial prompt 306 in a manner that improves the likelihood that the model 226 will generate an accurate support response that satisfies the user's requirements. (Mehrotra, 0046).
Claim 13 recite equivalent claim limitations as claim 3 above, except that they set forth the claimed invention as a system as such they are rejected for the same reasons as applied hereinabove.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Pathak et. al (US 20240394477 A1): see Abst., paragraphs 0044, 0062.
Siebel et al. (US 20240202221 A1): paragraph 0123
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DENNIS TRUONG whose telephone number is (571)270-3157. The examiner can normally be reached Monday - Friday 7:00 am - 3:30 pm PT.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Neveen Abel-Jail can be reached at 571-270-0474. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/DENNIS TRUONG/Primary Examiner, Art Unit 2152 03/17/2026