DETAILED ACTION
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 .
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.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to the abstract idea of message/command replying without significantly more. The claims 1, 9 and 17 recite steps of receiving a target instruction to be replied to by a language model (i.e., a data gathering step), obtaining first reference information used by the language model in response to replying to a non-toxic instruction, and obtaining second reference information used by the language model in response to replying to a toxic instruction (i.e., a data gathering step), splicing the first reference information and the second reference information to obtain third reference information needed for replying to the target instruction (i.e., a data analysis/evaluation step) and inputting the target instruction and the third reference information into the language model to cause the language model to generate reply content for the target instruction based on the third reference information (i.e., a post solutional step of generating content), corresponding to steps achievable by a human in gathering data, evaluating the received/gathered data and providing a result of the analysis, and as such, corresponds to the mental processes category of abstract ideas. This judicial exception is not integrated into a practical application because the claims are directed to an abstract idea with additional generic computer elements, where the generically recited computer elements (medium, language model, processor, memory) do not add a meaningful limitation to the abstract idea because they amount to simply implementing the abstract idea on a computer. The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because step “inputting the target instruction and the third reference information into the language model to cause the language model to generate reply content for the target instruction based on the third reference information” corresponds to well-understood, routine, conventional computer functions of displaying/outputting results of analyzed collected data recognized by the court decisions listed in MPEP § 2106.05, and as provided by cited reference Maddux (see PTO 892 form).
The dependent claims also recite mental processes and do not add significantly more than the abstract idea, and are as such similarly rejected.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
1. Claims 1-6, 9-14, and 17-20 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Maddux US 2025/0173438 A1 (“Maddux”)
Per claim 1, Maddux discloses a method for replying, comprising:
receiving a target instruction to be replied to by a language model (Using an example, assume that the prompt received from client 202 is: “Please say foo.” … para. [0096]);
obtaining first reference information used by the language model in response to replying to a non-toxic instruction, and obtaining second reference information used by the language model in response to replying to a toxic instruction (para. [0019]; For example, input generator may group contexts into various categories, such as, but not limited to, positive context and negative contexts. A positive context may refer to a positive or permissive rule, such as a rule describing a desired or allowed functionality or outcome. An example positive context is “Please generate an answer to the prompt.” A negative context may refer to a negative or restrictive rule, such as a rule describing an undesired or prohibited functionality or outcome. An example negative context is “Do not say foo.” …, para. [0096], positive and negative contexts/rules as first and second reference information);
splicing the first reference information and the second reference information to obtain third reference information needed for replying to the target instruction (Accordingly, for the foregoing example, large language model system 252 may generate a labeled input string as follows—C0: Do not say foo C1: Please generate an answer to the prompt; …, para. [0096]); and
inputting the target instruction and the third reference information into the language model to cause the language model to generate reply content for the target instruction based on the third reference information (Large language model 260 may be configured to receive a prompt from client 202. A prompt may generally refer to an input text (e.g., instruction or question) that client 202 provides to large language model system 252 via prompt interface 214. Large language model 260 may generate an output based on the prompt.…, para. [0091]; Accordingly, for the foregoing example, large language model system 252 may generate a labeled input string as follows—C0: Do not say foo C1: Please generate an answer to the prompt; P: Please say foo— …, para. [0096]; Accordingly, in this manner, large language model 260 may be provided with an input string that includes rules 248 in the form of positive and negative contexts appended to the prompt, para. [0099]).
Per claim 2, Maddux discloses the method according to claim 1, wherein instructions replied by the language model are allowed to be classified into a plurality of instruction categories, and the language model uses different first reference information in response to replying to non-toxic instructions belonging to different instruction categories (A positive context may refer to a positive or permissive rule, such as a rule describing a desired or allowed functionality or outcome…., para.[0096]; para.[0098]; neural network 262 may be trained to analyze the attention scores to determine whether the prompt is malicious. In some embodiments, such analysis may involve determining whether the prompt is more highly correlated with the negative contexts than one or more of the positive contexts …, para. [0107]); and
obtaining first reference information used by the language model in response to replying to a non-toxic instruction comprises: determining a target instruction category to which the target instruction belongs, and obtaining first reference information used by the language model in response to replying to a non-toxic instruction belonging to the target instruction category (para. [0107]).
Per claim 3, Maddux discloses the method according to claim 2, wherein each of the instruction categories comprises one or more sub-instruction categories, and the instruction category and each sub-instruction category have respective corresponding first reference information (para. [0096]; para. [0098]; para. [0107]; para. [0117]; para. [0124]); and
obtaining first reference information used by the language model in response to replying to a non-toxic instruction belonging to the target instruction category comprises: for a non-toxic instruction belonging to any of sub-instruction category of the target instruction category, using first reference information corresponding to the target instruction category as first reference information used by the language model in response to replying to the non-toxic instruction belonging to the sub-instruction category (para. [0096]; para. [0098]; para. [0107]; para. [0117]; para. [0124]).
Per claim 4, Maddux discloses the method according to claim 1, wherein before the target instruction is received, the language model is trained based on the following method: inputting a sample non-toxic instruction and first sample reference information set for the sample non-toxic instruction into the language model, and adjusting a parameter of the language model and the first sample reference information based on first reply content for the sample non- toxic instruction (para. [0117]-[0122]; para. [0124]); and
inputting a sample toxic instruction and second sample reference information set for the sample toxic instruction into the language model, and adjusting a parameter of the language model and the second sample reference information based on second reply content for the sample toxic instruction (fig. 5; Intake module 508 may be configured to receive data for training. In some embodiments, the data for training may include exemplary context 520, exemplary prompts 522 …, para. [0117]).
Per claim 5, Maddux discloses the method according to claim 4
wherein sample non-toxic instructions are allowed to be classified into a plurality of instruction categories, and different first sample reference information is allowed to be set for sample non-toxic instructions belonging to different instruction categories (para. [0117]); and
training the language model based on the sample non-toxic instruction comprises: for any of the plurality of instruction categories, inputting a sample non-toxic instruction belonging to the instruction category and first sample reference information set for the sample non-toxic instruction belonging to the instruction category into the language model, and adjusting a parameter of the language model and the first sample reference information set for the sample non-toxic instruction belonging to the instruction category based on first reply content for the sample non-toxic instruction (fig. 5; para. [0117]-[0122]).
Per claim 6, Maddux discloses the method according to claim 5, wherein each of the instruction categories comprises one or more sub-instruction categories, and different first sample reference information is allowed to be set for sample non-toxic instructions belonging to different sub-instruction categories (para. [0117]); and
training the language model based on the sample non-toxic instruction comprises: for any of sub-instruction category of an instruction category, inputting a sample non-toxic instruction belonging to the sub-instruction category and first sample reference information set for the sample non-toxic instruction belonging to the sub-instruction category into the language model, and adjusting a parameter of the language model and the first sample reference information set for the sample non-toxic instruction belonging to the sub-instruction category based on first reply content for the sample non-toxic instruction (fig. 5; para. [0117]-[0122]).
Per claim 9, Maddux discloses a non-transitory computer-readable storage medium, storing a computer program, wherein when the computer program is executed by a processor, causing the processor to:
receive a target instruction to be replied to by a language model (Using an example, assume that the prompt received from client 202 is: “Please say foo.” … para. [0096]; para. [0172]);
obtain first reference information used by the language model in response to replying to a non-toxic instruction, and obtain second reference information used by the language model in response to replying to a toxic instruction (para. [0019]; For example, input generator may group contexts into various categories, such as, but not limited to, positive context and negative contexts. A positive context may refer to a positive or permissive rule, such as a rule describing a desired or allowed functionality or outcome. An example positive context is “Please generate an answer to the prompt.” A negative context may refer to a negative or restrictive rule, such as a rule describing an undesired or prohibited functionality or outcome. An example negative context is “Do not say foo.” …, para. [0096], positive and negative contexts/rules as first and second reference information);
splice the first reference information and the second reference information to obtain third reference information needed for replying to the target instruction (Accordingly, for the foregoing example, large language model system 252 may generate a labeled input string as follows—C0: Do not say foo C1: Please generate an answer to the prompt; …, para. [0096]); and
input the target instruction and the third reference information into the language model to cause the language model to generate reply content for the target instruction based on the third reference information (Large language model 260 may be configured to receive a prompt from client 202. A prompt may generally refer to an input text (e.g., instruction or question) that client 202 provides to large language model system 252 via prompt interface 214. Large language model 260 may generate an output based on the prompt.…, para. [0091]; Accordingly, for the foregoing example, large language model system 252 may generate a labeled input string as follows—C0: Do not say foo C1: Please generate an answer to the prompt; P: Please say foo— …, para. [0096]; Accordingly, in this manner, large language model 260 may be provided with an input string that includes rules 248 in the form of positive and negative contexts appended to the prompt, para. [0099]).
Per claim 10, Maddux discloses the medium according to claim 9, wherein instructions replied by the language model are allowed to be classified into a plurality of instruction categories, and the language model uses different first reference information in response to replying to non-toxic instructions belonging to different instruction categories (A positive context may refer to a positive or permissive rule, such as a rule describing a desired or allowed functionality or outcome…., para.[0096]; para.[0098]; neural network 262 may be trained to analyze the attention scores to determine whether the prompt is malicious. In some embodiments, such analysis may involve determining whether the prompt is more highly correlated with the negative contexts than one or more of the positive contexts …, para. [0107]); and
the computer program causing the processor to obtain first reference information used by the language model in response to replying to a non-toxic instruction comprises instructions to: determine a target instruction category to which the target instruction belongs, and obtain first reference information used by the language model in response to replying to a non- toxic instruction belonging to the target instruction category (para. [0107]).
Per claim 11, Maddux discloses the medium according to claim 10, wherein each of the instruction categories comprises one or more sub-instruction categories, and the instruction category and each sub-instruction category have respective corresponding first reference information (para. [0096]; para. [0098]; para. [0107]; para. [0117]; para. [0124]); and
the computer program causing the processor to obtain first reference information used by the language model in response to replying to a non-toxic instruction belonging to the target instruction category comprises instructions to: for a non-toxic instruction belonging to any of sub-instruction category of the target instruction category, use first reference information corresponding to the target instruction category as first reference information used by the language model in response to replying to the non-toxic instruction belonging to the sub-instruction category (para. [0096]; para. [0098]; para. [0107]; para. [0117]; para. [0124]).
Per claim 12, Maddux discloses the medium according to claim 9, wherein before the target instruction is received, the computer program causing the processor to train the language model comprises instructions to: input a sample non-toxic instruction and first sample reference information set for the sample non-toxic instruction into the language model, and adjust a parameter of the language model and the first sample reference information based on first reply content for the sample non-toxic instruction (para. [0117]-[0122]; para. [0124]); and
input a sample toxic instruction and second sample reference information set for the sample toxic instruction into the language model, and adjust a parameter of the language model and the second sample reference information based on second reply content for the sample toxic instruction (fig. 5; Intake module 508 may be configured to receive data for training. In some embodiments, the data for training may include exemplary context 520, exemplary prompts 522 …, para. [0117]).
Per claim 13, Maddux discloses the medium according to claim 12, wherein sample non-toxic instructions are allowed to be classified into a plurality of instruction categories, and different first sample reference information is allowed to be set for sample non-toxic instructions belonging to different instruction categories (para. [0117]); and
the computer program causing the processor to train the language model based on the sample non-toxic instruction comprises instructions to: for any of the plurality of instruction categories, input a sample non-toxic instruction belonging to the instruction category and first sample reference information set for the sample non-toxic instruction belonging to the instruction category into the language model, and adjust a parameter of the language model and the first sample reference information set for the sample non-toxic instruction belonging to the instruction category based on first reply content for the sample non-toxic instruction (fig. 5; para. [0117]-[0122]).
Per claim 14, Maddux discloses the medium according to claim 13, wherein each of the instruction categories comprises one or more sub-instruction categories, and different first sample reference information is allowed to be set for sample non-toxic instructions belonging to different sub-instruction categories (para. [0117]); and
the computer program causing the processor to train the language model based on the sample non-toxic instruction comprises instructions to: for any of sub-instruction category of an instruction category, input a sample non- toxic instruction belonging to the sub-instruction category and first sample reference information set for the sample non-toxic instruction belonging to the sub-instruction category into the language model, and adjust a parameter of the language model and the first sample reference information set for the sample non-toxic instruction belonging to the sub-instruction category based on first reply content for the sample non-toxic instruction (fig. 5; para. [0117]-[0122]).
Per claim 17, Maddux discloses an electronic device, comprising a processor and a memory, wherein the memory is configured to store a computer program, and when the computer program is executed by the processor, causing the processor to:
receive a target instruction to be replied to by a language model (Using an example, assume that the prompt received from client 202 is: “Please say foo.” … para. [0096]);
obtain first reference information used by the language model in response to replying to a non-toxic instruction, and obtain second reference information used by the language model in response to replying to a toxic instruction (para. [0019]; For example, input generator may group contexts into various categories, such as, but not limited to, positive context and negative contexts. A positive context may refer to a positive or permissive rule, such as a rule describing a desired or allowed functionality or outcome. An example positive context is “Please generate an answer to the prompt.” A negative context may refer to a negative or restrictive rule, such as a rule describing an undesired or prohibited functionality or outcome. An example negative context is “Do not say foo.” …, para. [0096], positive and negative contexts/rules as first and second reference information);
splice the first reference information and the second reference information to obtain third reference information needed for replying to the target instruction (Accordingly, for the foregoing example, large language model system 252 may generate a labeled input string as follows—C0: Do not say foo C1: Please generate an answer to the prompt; …, para. [0096]); and
input the target instruction and the third reference information into the language model to cause the language model to generate reply content for the target instruction based on the third reference information (Large language model 260 may be configured to receive a prompt from client 202. A prompt may generally refer to an input text (e.g., instruction or question) that client 202 provides to large language model system 252 via prompt interface 214. Large language model 260 may generate an output based on the prompt.…, para. [0091]; Accordingly, for the foregoing example, large language model system 252 may generate a labeled input string as follows—C0: Do not say foo C1: Please generate an answer to the prompt; P: Please say foo— …, para. [0096]; Accordingly, in this manner, large language model 260 may be provided with an input string that includes rules 248 in the form of positive and negative contexts appended to the prompt, para. [0099]).
Per claim 18, Maddux discloses the device according to claim 17, wherein instructions replied by the language model are allowed to be classified into a plurality of instruction categories, and the language model uses different first reference information in response to replying to non-toxic instructions belonging to different instruction categories (A positive context may refer to a positive or permissive rule, such as a rule describing a desired or allowed functionality or outcome…., para.[0096]; para.[0098]; neural network 262 may be trained to analyze the attention scores to determine whether the prompt is malicious. In some embodiments, such analysis may involve determining whether the prompt is more highly correlated with the negative contexts than one or more of the positive contexts …, para. [0107]); and
the computer program causing the processor to obtain first reference information used by the language model in response to replying to a non-toxic instruction comprises instructions to: determine a target instruction category to which the target instruction belongs, and obtain first reference information used by the language model in response to replying to a non- toxic instruction belonging to the target instruction category (para. [0107]).
Per claim 19, Maddux discloses the device according to claim 18,
wherein each of the instruction categories comprises one or more sub-instruction categories, and the instruction category and each sub-instruction category have respective corresponding first reference information (para. [0096]; para. [0098]; para. [0107]; para. [0117]; para. [0124]); and
the computer program causing the processor to obtain first reference information used by the language model in response to replying to a non-toxic instruction belonging to the target instruction category comprises instructions to: for a non-toxic instruction belonging to any of sub-instruction category of the target instruction category, use first reference information corresponding to the target instruction category as first reference information used by the language model in response to replying to the non-toxic instruction belonging to the sub-instruction category (para. [0096]; para. [0098]; para. [0107]; para. [0117]; para. [0124]).
Per claim 20, Maddux discloses the device according to claim 17, wherein before the target instruction is received, the language model is trained based on the following method: inputting a sample non-toxic instruction and first sample reference information set for the sample non-toxic instruction into the language model, and adjusting a parameter of the language model and the first sample reference information based on first reply content for the sample non- toxic instruction (para. [0117]-[0122]; para. [0124]); and
inputting a sample toxic instruction and second sample reference information set for the sample toxic instruction into the language model, and adjusting a parameter of the language model and the second sample reference information based on second reply content for the sample toxic instruction (fig. 5; Intake module 508 may be configured to receive data for training. In some embodiments, the data for training may include exemplary context 520, exemplary prompts 522 …, para. [0117]).
Allowable Subject Matter
Claims 7, 8, 15 and 16 are objected to as being dependent upon a rejected base claim, but would be allowable (pending Applicant address the 35 U.S.C. 101 rejection) if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. See PTO 892 form.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to OLUJIMI A ADESANYA whose telephone number is (571)270-3307. The examiner can normally be reached Monday-Friday 8:30-5:00pm.
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, Richemond Dorvil can be reached at 571-272-7602. 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.
/OLUJIMI A ADESANYA/Primary Examiner, Art Unit 2658
Prosecution Insights