【翻译前 - 英文原文】

================================================================================
1 ABSTRACT
================================================================================
While large language models (LLMs) have demonstrated impressive performance
across tasks in language understanding and interactive decision making, their
abilities for reasoning (e.g. chain-of-thought prompting) and acting (e.g. action
plan generation) have primarily been studied as separate topics. In this paper, we
explore the use of LLMs to generate both reasoning traces and task-speciﬁc actions
in an interleaved manner, allowing for greater synergy between the two: reasoning
traces help the model induce, track, and update action plans as well as handle
exceptions, while actions allow it to interface with and gather additional information
from external sources such as knowledge bases or environments. We apply our
approach, named ReAct, to a diverse set of language and decision making tasks
and demonstrate its effectiveness over state-of-the-art baselines in addition to
improved human interpretability and trustworthiness. Concretely, on question
answering (HotpotQA) and fact veriﬁcation (Fever), ReAct overcomes prevalent
issues of hallucination and error propagation in chain-of-thought reasoning by
interacting with a simple Wikipedia API, and generating human-like task-solving
trajectories that are more interpretable than baselines without reasoning traces.
Furthermore, on two interactive decision making benchmarks (ALFWorld and
WebShop), ReAct outperforms imitation and reinforcement learning methods by
an absolute success rate of 34% and 10% respectively, while being prompted with
only one or two in-context examples.

================================================================================
2 INTRODUCTION
================================================================================
A unique feature of human intelligence is the ability to seamlessly combine task-oriented actions with
verbal reasoning (or inner speech, Alderson-Day & Fernyhough, 2015), which has been theorized to
play an important role in human cognition for enabling self-regulation or strategization (Vygotsky,
1987; Luria, 1965; Fernyhough, 2010) and maintaining a working memory (Baddeley, 1992). Con-
sider the example of cooking up a dish in the kitchen. Between any two speciﬁc actions, we may
reason in language in order to track progress (“now that everything is cut, I should heat up the pot of
water”), to handle exceptions or adjust the plan according to the situation (“I don’t have salt, so let
me use soy sauce and pepper instead”), and to realize when external information is needed (“how do
I prepare dough? Let me search on the Internet”). We may also act (open a cookbook to read the
recipe, open the fridge, check ingredients) to support the reasoning and to answer questions (“What
dish can I make right now?”). This tight synergy between “acting” and “reasoning” allows humans
to learn new tasks quickly and perform robust decision making or reasoning, even under previously
unseen circumstances or facing information uncertainties.
Recent results have hinted at the possibility of combining verbal reasoning with interactive decision
making in autonomous systems. On one hand, properly prompted large language models (LLMs)
have demonstrated emergent capabilities to carry out several steps of reasoning traces to derive
∗Work during Google internship. Projet page with code: https://react-lm.github.io/.
arXiv:2210.03629v3  [cs.CL]  10 Mar 2023

--- Page 2 ---
$FW7KLQN>)LUVW,QHHGWRILQGDSHSSHUVKDNHUPRUH 
OLNHO\WRDSSHDULQFDELQHWVFRXQWHUWRSV
@
$FW*RWRFDELQHW 
2EV2QWKHFDELQHW\RXVHHDYDVH 
+HUHWKHDJHQWJRWRFDELQHWWKHQFDELQHWWKHQFRXQWHUWRSDQG
$FW*RWRFRXQWHUWRS 
2EV2QWKHFRXQWHUWRS\RXVHHDDSSOHDEUHDG

DSHSSHUVKDNHUDQGDYDVH
$FW7DNHSHSSHUVKDNHUIURPFRXQWHUWRS 
2EV<RXSLFNXSWKHSHSSHUVKDNHUIURPWKHFRXQWHUWRS

$FW7KLQN>1RZ,ILQGDSHSSHUVKDNHU1H[W,QHHGWR

SXWLWLQRQGUDZHU @
$FW*RWRGUDZHU 
2EV'UDZHULVFORVHG 
$FW2SHQGUDZHU 
2EV<RXRSHQ'UDZHUª
$FW3XWSHSSHUVKDNHULQRQGUDZHU 
2EV<RXSXWSHSSHUVKDNHULQRQWKHGUDZHU
D6WDQGDUG
$QVZHUL3RG
E&R75HDVRQ2QO\
7KRXJKW/HW
VWKLQNVWHS 
E\VWHS $SSOH5HPRWHZDV 
RULJLQDOO\GHVLJQHGWR 
LQWHUDFWZLWK$SSOH79 
$SSOH79FDQEHFRQWUROOHG 
E\L3KRQHL3DGDQGL3RG 
7RXFK6RWKHDQVZHULV 
L3KRQHL3DGDQGL3RG 
7RXFK
$QVZHUL3KRQHL3DGL3RG 
7RXFK
F$FW2QO\
$FW6HDUFK>$SSOH5HPRWH@ 
2EV7KH$SSOH5HPRWHLV D
UHPRWHFRQWUROª
$FW6HDUFK>)URQW5RZ @
2EV&RXOGQRWILQG>)URQW 
5RZ@6LPLODUª
$FW6HDUFK>)URQW5RZ 
VRIWZDUH @
2EV)URQW5RZLVD 
GLVFRQWLQXHGPHGLDFHQWHU 
VRIWZDUHª
$FW)LQLVK>\HV@
G5H$FW5HDVRQ$FW
7KRXJKW,QHHGWRVHDUFK$SSOH5HPRWHDQGILQGWKH 
SURJUDPLWZDVRULJLQDOO\GHVLJQHGWRLQWHUDFWZLWK
$FW6HDUFK>$SSOH5HPRWH@
2EV7KH$SSOH5HPRWHLVDUHPRWHFRQWUROLQWURGXFHGLQ

2FWREHUE\$SSOHªRULJLQDOO\GHVLJQHGWRFRQWUROWKH

)URQW5RZPHGLDFHQWHU SURJUDPª
7KRXJKW$SSOH5HPRWHZDVRULJLQDOO\GHVLJQHGWRFRQWURO

WKH)URQW5RZPHGLDFHQWHUSURJUDP,QHHGWRVHDUFK
)URQW
5RZQH[WDQGILQGZKDWRWKHUGHYLFHFDQFRQWUROLW
$FW6HDUFK>)URQW5RZ@
2EV&RXOGQRWILQG>)URQW5RZ@6LPLODU>
)URQW5RZ

6HDWWR(DUWK

)URQW5RZ0RWRUVSRUWV

 )URQW5RZ 
VRIWZDUH 
ª@
7KRXJKW)URQW5RZLVQRWIRXQG,QHHGWRVHDUFK )URQW
5RZVRIWZDUH 
$FW6HDUFK>)URQW5RZVRIWZDUH @
2EV)URQW5RZLVDGLVFRQWLQXHGPHGLDFHQWHUVRIWZDUHª
7KRXJKW)URQW5RZVRIWZDUHLVFRQWUROOHGE\DQ$SSOH

5HPRWHRUWKHNH\ERDUGIXQFWLRQNH\V 6RWKHDQVZHULV 
NH\ERDUGIXQFWLRQNH\V
$FW)LQLVK>NH\ERDUGIXQFWLRQNH\V @
Ҽ
4XHVWLRQ$VLGHIURPWKH$SSOH5HPRWHZKDWRWKHUGHYLFH

FDQFRQWUROWKHSURJUDP$SSOH5HPRWHZDVRULJLQDOO\

GHVLJQHGWRLQWHUDFWZLWK"
D$FW2QO\
$FW*RWRGUDZHU 
2EV7KHGUDZHULVFORVHG 
$FW2SHQGUDZHU 
2EV<RXRSHQWKHGUDZHU7KHGUDZHULVRSHQ,QLW

\RXVHHDGLVKVSRQJHDQGDVSRRQ

$FW*RWRVLQNEDVLQ 
2EV2QWKHVLQNEDVLQ\RXVHHDGLVKVSRQJHD

VSDWXODDQGDVSRRQ 
$FW7DNHSHSSHUVKDNHUIURPVLQNEDVLQ 
2EV1RWKLQJKDSSHQV 
$FW7DNHSHSSHUVKDNHUIURPVLQNEDVLQ 
2EV1RWKLQJKDSSHQV 
E5H$FW5HDVRQ$FW
Ҽ
<RXDUHLQWKHPLGGOHRIDURRP/RRNLQJTXLFNO\DURXQG

\RX\RXVHHDFDELQHWDFDELQHWDFRIIHHPDFKLQH

DFRXQWHUWRSDVWRYHEXUQHUDQGDWRDVWHU

<RXUWDVNLVWR3XWVRPHSHSSHUVKDNHURQDGUDZHU
$OI:RUOG
+RWVSRW4$
Figure 1: (1) Comparison of 4 prompting methods, (a) Standard, (b) Chain-of-thought (CoT,
Reason Only), (c) Act-only, and (d) ReAct (Reason+Act), solving a HotpotQA (Yang et al., 2018)
question. (2) Comparison of (a) Act-only and (b) ReAct prompting to solve an AlfWorld (Shridhar
et al., 2020b) game. In both domains, we omit in-context examples in the prompt, and only show task
solving trajectories generated by the model (Act, Thought) and the environment (Obs).
answers from questions in arithmetic, commonsense, and symbolic reasoning tasks (Wei et al.,
2022). However, this “chain-of-thought” reasoning is a static black box, in that the model uses
its own internal representations to generate thoughts and is not grounded in the external world,
which limits its ability to reason reactively or update its knowledge. This can lead to issues like fact
hallucination and error propagation over the reasoning process (Figure 1 (1b)). On the other hand,
recent work has explored the use of pre-trained language models for planning and acting in interactive
environments (Ahn et al., 2022; Nakano et al., 2021; Yao et al., 2020; Huang et al., 2022a), with
a focus on predicting actions via language priors. These approaches usually convert multi-modal
observations into text, use a language model to generate domain-speciﬁc actions or plans, and then
use a controller to choose or execute them. However, they do not employ language models to reason
abstractly about high-level goals or maintain a working memory to support acting, barring Huang
et al. (2022b) who perform a limited form of verbal reasoning to reiterate spatial facts about the
current state. Beyond such simple embodied tasks to interact with a few blocks, there have not been
studies on how reasoning and acting can be combined in a synergistic manner for general task solving,
and if such a combination can bring systematic beneﬁts compared to reasoning or acting alone.
In this work, we present ReAct, a general paradigm to combine reasoning and acting with language
models for solving diverse language reasoning and decision making tasks (Figure 1). ReAct
prompts LLMs to generate both verbal reasoning traces and actions pertaining to a task in an
interleaved manner, which allows the model to perform dynamic reasoning to create, maintain, and
adjust high-level plans for acting (reason to act), while also interact with the external environments
(e.g. Wikipedia) to incorporate additional information into reasoning (act to reason).

--- Page 3 ---
We conduct empirical evaluations of ReAct and state-of-the-art baselines on four diverse benchmarks:
question answering (HotPotQA, Yang et al., 2018), fact veriﬁcation (Fever, Thorne et al., 2018),
text-based game (ALFWorld, Shridhar et al., 2020b), and webpage navigation (WebShop, Yao
et al., 2022). For HotPotQA and Fever, with access to a Wikipedia API that the model can interact
with, ReAct outperforms vanilla action generation models while being competitive with chain-of-
thought reasoning (CoT) (Wei et al., 2022). The best approach overall is a combination of ReAct
and CoT that allows for the use of both internal knowledge and externally obtained information
during reasoning. On ALFWorld and WebShop, two or even one-shot ReAct prompting is able
to outperform imitation or reinforcement learning methods trained with 103 ∼105 task instances,
with an absolute improvement of 34% and 10% in success rates respectively. We also demonstrate
the importance of sparse, versatile reasoning in decision making by showing consistent advantages
over controlled baselines with actions only. Besides general applicability and performance boost,
the combination of reasoning and acting also contributes to model interpretability, trustworthiness,
and diagnosability across all domains, as humans can readily distinguish information from model’s
internal knowledge versus external environments, as well as inspect reasoning traces to understand
the decision basis of model actions.
To summarize, our key contributions are the following: (1) we introduce ReAct, a novel prompt-
based paradigm to synergize reasoning and acting in language models for general task solving; (2) we
perform extensive experiments across diverse benchmarks to showcase the advantage of ReAct in a
few-shot learning setup over prior approaches that perform either reasoning or action generation in
isolation; (3) we present systematic ablations and analysis to understand the importance of acting in
reasoning tasks, and reasoning in interactive tasks; (4) we analyze the limitations of ReAct under the
prompting setup (i.e. limited support of reasoning and acting behaviors), and perform initial ﬁnetuning
experiments showing the potential of ReAct to improve with additional training data. Scaling up
ReAct to train and operate on more tasks and combining it with complementary paradigms like
reinforcement learning could further unlock the potential of large language models.
REAC T: SYNERGIZING REASONING + ACTING
Consider a general setup of an agent interacting with an environment for task solving. At time
step t, an agent receives an observation ot ∈O from the environment and takes an action at ∈A
following some policy π(at|ct), where ct = (o1, a1, · · · , ot−1, at−1, ot) is the context to the agent.
Learning a policy is challenging when the mapping ct 7→at is highly implicit and requires extensive
computation. For example, the agent shown in Figure 1(1c) is unable to generate the correct ﬁnal
action (Act 4) to ﬁnish the QA task as it requires complex reasoning over the trajectory context
(Question, Act 1-3, Obs 1-3). Similarly, the agent shown in Figure 1(2a) fails to comprehend from the
context that sinkbasin 1 does not contain peppershaker 1, thus keep producing hallucinating actions.
The idea of ReAct is simple: we augment the agent’s action space to ˆ
A = A ∪L, where L is the
space of language. An action ˆat ∈L in the language space, which we will refer to as a thought or a
reasoning trace, does not affect the external environment, thus leading to no observation feedback.
Instead, a thought ˆat aims to compose useful information by reasoning over the current context ct,
and update the context ct+1 = (ct, ˆat) to support future reasoning or acting. As shown in Figure 1,
there could be various types of useful thoughts, e.g. decomposing task goals and create action plans
(2b, Act 1; 1d, Thought 1), injecting commonsense knowledge relevant to task solving (2b, Act 1),
extracting important parts from observations (1d, Thought2, 4), track progress and transit action plans
(2b, Act 8), handle exceptions and adjust action plans (1d, Thought 3), and so on.
However, as the language space L is unlimited, learning in this augmented action space is difﬁcult
and requires strong language priors. In this paper, we mainly focus on the setup where a frozen
large language model, PaLM-540B (Chowdhery et al., 2022)1, is prompted with few-shot in-context
examples to generate both domain-speciﬁc actions and free-form language thoughts for task solving
(Figure 1 (1d), (2b)). Each in-context example is a human trajectory of actions, thoughts, and
environment observations to solve a task instance (see Appendix C). For the tasks where reasoning is
of primary importance (Figure 1(1)), we alternate the generation of thoughts and actions so that the
task-solving trajectory consists of multiple thought-action-observation steps. In contrast, for decision
making tasks that potentially involve a large number of actions (Figure 1(2)), thoughts only need to
1We show some GPT-3 (Brown et al., 2020) results in Appendix A.1, which outperforms PaLM-540B.

--- Page 4 ---
appear sparsely in the most relevant positions of a trajectory, so we let the language model decide the
asynchronous occurrence of thoughts and actions for itself.
Since decision making and reasoning capabilities are integrated into a large language model, ReAct
enjoys several unique features: A) Intuitive and easy to design: Designing ReAct prompts is
straightforward as human annotators just type down their thoughts in language on top of their actions
taken. No ad-hoc format choice, thought design, or example selection is used in this paper. We detail
prompt design for each task in Sections 3 and 4. B) General and ﬂexible: Due to the ﬂexible thought
space and thought-action occurrence format, ReAct works for diverse tasks with distinct action
spaces and reasoning needs, including but not limited to QA, fact veriﬁcation, text game, and web
navigation. C) Performant and robust: ReAct shows strong generalization to new task instances
while learning solely from one to six in-context examples, consistently outperforming baselines with
only reasoning or acting across different domains. We also show in Section 3 additional beneﬁts
when ﬁnetuning is enabled, and in Section 4 how ReAct performance is robust to prompt selections.
D) Human aligned and controllable: ReAct promises an interpretable sequential decision making
and reasoning process where humans can easily inspect reasoning and factual correctness. Moreover,
humans can also control or correct the agent behavior on the go by thought editing, as shown in
Figure 5 in Section 4.
KNOWLEDGE-INTENSIVE REASONING TASKS
We begin with knowledge-intensive reasoning tasks like multi-hop question answering and fact
veriﬁcation. As shown in Figure 1(1d), by interacting with a Wikipedia API, ReAct is able to
retrieve information to support reasoning, while also use reasoning to target what to retrieve next,
demonstrating a synergy of reasoning and acting.
3.1
SETUP
Domains
We consider two datasets challenging knowledge retrieval and reasoning: (1) Hot-
PotQA (Yang et al., 2018), a multi-hop question answering benchmark that requires reasoning
over two or more Wikipedia passages, and (2) FEVER (Thorne et al., 2018), a fact veriﬁcation
benchmark where each claim is annotated SUPPORTS, REFUTES, or NOT ENOUGH INFO, based
on if there exists a Wikipedia passage to verify the claim. In this work, we operate in a question-only
setup for both tasks, where models only receive the question/claim as input without access to support
paragraphs, and have to rely on their internal knowledge or retrieve knowledge via interacting with
an external environment to support reasoning.
Action Space
We design a simple Wikipedia web API with three types of actions to support
interactive information retrieval: (1) search[entity], which returns the ﬁrst 5 sentences from
the corresponding entity wiki page if it exists, or else suggests top-5 similar entities from the
Wikipedia search engine, (2) lookup[string], which would return the next sentence in the page
containing string, simulating Ctrl+F functionality on the browser. (3) finish[answer], which
would ﬁnish the current task with answer. We note that this action space mostly can only retrieve a
small part of a passage based on exact passage name, which is signiﬁcantly weaker than state-of-the-
art lexical or neural retrievers. The purpose is to simulate how humans would interact with Wikipedia,
and force models to retrieve via explicit reasoning in language.
3.2

================================================================================
3 METHOD
================================================================================
S
ReAct Prompting
For HotpotQA and Fever, we randomly select 6 and 3 cases2 from the training
set and manually compose ReAct-format trajectories to use as few-shot exemplars in the prompts.
Similar to Figure 1(d), each trajectory consists of multiple thought-action-observation steps (i.e. dense
thought), where free-form thoughts are used for various purposes. Speciﬁcally, we use a combination
of thoughts that decompose questions (“I need to search x, ﬁnd y, then ﬁnd z”), extract information
from Wikipedia observations (“x was started in 1844”, “The paragraph does not tell x”), perform
commonsense (“x is not y, so z must instead be...”) or arithmetic reasoning (“1844 < 1989”), guide
2We ﬁnd more examples do not improve performance.

--- Page 5 ---
Prompt Methoda
HotpotQA
Fever
(EM)
(Acc)
Standard
28.7
57.1
CoT (Wei et al., 2022)
29.4
56.3
CoT-SC (Wang et al., 2022a)
33.4
60.4
Act
25.7
58.9
ReAct
27.4
60.9
CoT-SC →ReAct
34.2
64.6
ReAct→CoT-SC
35.1
62.0
Supervised SoTAb
67.5
89.5
Table 1: PaLM-540B prompting results on
HotpotQA and Fever.
aHotpotQA EM is 27.1, 28.9, 33.8 for Standard, CoT,
CoT-SC in Wang et al. (2022b).
b(Zhu et al., 2021; Lewis et al., 2020)
0
20
#CoT-SC trials
26
28
30
32
34
HotpotQA EM
0
20
#CoT-SC trials
47.5
50.0
52.5
55.0
57.5
60.0
62.5
65.0
Fever Acc
Method
CoT-SC -> ReAct
ReAct -> CoT-SC
CoT-SC
ReAct
CoT
Figure 2: PaLM-540B prompting results with respect to
number of CoT-SC samples used.
search reformulation (“maybe I can search/look up x instead”), and synthesize the ﬁnal answer (“...so
the answer is x”). See Appendix C for more details.
Baselines
We systematically ablate ReAct trajectories to build prompts for multiple baselines (with
formats as Figure 1(1a-1c)): (a) Standard prompting (Standard), which removes all thoughts,
actions, observations in ReAct trajectories. (b) Chain-of-thought prompting (CoT) (Wei et al.,
2022), which removes actions and observations and serve as a reasoning-only baseline. We also
build a self-consistency baseline (CoT-SC) (Wang et al., 2022a;b) by sampling 21 CoT trajectories
with decoding temperature 0.7 during inference and adopting the majority answer, which is found to
consistently boost performance over CoT. (c) Acting-only prompt (Act), which removes thoughts
in ReAct trajectories, loosely resembling how WebGPT (Nakano et al., 2021) interacts with the
Internet to answer questions, though it operates on a different task and action space, and uses imitation
and reinforcement learning instead of prompting.
Combining Internal and External Knowledge
As will be detail in Section 3.3, we observe that
the problem solving process demonstrated by ReAct is more factual and grounded, whereas CoT
is more accurate in formulating reasoning structure but can easily suffer from hallucinated facts
or thoughts. We therefore propose to incorporate ReAct and CoT-SC, and let the model decide
when to switch to the other method based on the following heuristics: A) ReAct →CoT-SC: when
ReAct fails to return an answer within given steps, back off to CoT-SC. We set 7 and 5 steps for
HotpotQA and FEVER respectively as we ﬁnd more steps will not improve ReAct performance3.
B) CoT-SC →ReAct: when the majority answer among n CoT-SC samples occurs less than n/2
times (i.e. internal knowledge might not support the task conﬁdently), back off to ReAct.
Finetuning
Due to the challenge of manually annotating reasoning traces and actions at scale,
we consider a bootstraping approach similar to Zelikman et al. (2022), using 3,000 trajectories
with correct answers generated by ReAct (also for other baselines) to ﬁnetune smaller language
models (PaLM-8/62B) to decode trajectories (all thoughts, actions, observations) conditioned on
input questions/claims. More details are in Appendix B.1.
3.3
RESULTS AND OBSERVATIONS
ReAct outperforms Act consistently
Table 1 shows HotpotQA and Fever results using PaLM-
540B as the base model with different prompting methods. We note that ReAct is better than Act
on both tasks, demonstrating the value of reasoning to guide acting, especially for synthesizing the
ﬁnal answer, as shown in Figure 1 (1c-d). Fine-tuning results 3 also conﬁrm the beneﬁt of reasoning
traces for more informed acting.
3Of all trajectories with correct ﬁnal answers, those with 7 steps on HotpotQA and 5 steps on FEVER only
take up 0.84% and 1.33% respectively.

--- Page 6 ---
Type
Deﬁnition
ReAct
CoT
Success
True positive
Correct reasoning trace and facts
94%
86%
False positive
Hallucinated reasoning trace or facts
6%
14%
Failure
Reasoning error
Wrong reasoning trace (including failing to recover from repetitive steps)
47%
16%
Search result error
Search return empty or does not contain useful information
23%
-
Hallucination
Hallucinated reasoning trace or facts
0%
56%
Label ambiguity
Right prediction but did not match the label precisely
29%
28%
Table 2: Types of success and failure modes of ReAct and CoT on HotpotQA, as well as their
percentages in randomly selected examples studied by human.
ReAct vs. CoT
On the other hand, ReAct outperforms CoT on Fever (60.9 vs. 56.3) and slightly
lags behind CoT on HotpotQA (27.4 vs. 29.4). Fever claims for SUPPORTS/REFUTES might only
differ by a slight amount (see Appendix D.1), so acting to retrieve accurate and up-to-date knowledge
is vital. To better understand the behavioral difference between ReAct and CoT on HotpotQA, we
randomly sampled 50 trajectories with correct and incorrect answers (judged by EM) from ReAct
and CoT respectively (thus 200 examples in total), and manually labeled their success and failure
modes in Table 2. Some key observations are as follows:
A) Hallucination is a serious problem for CoT, resulting in much higher false positive rate than
ReAct (14% vs. 6%) in success mode, and make up its major failure mode (56%). In contrast, the
problem solving trajectory of ReActis more grounded, fact-driven, and trustworthy, thanks to the
access of an external knowledge base.
B) While interleaving reasoning, action and observation steps improves ReAct’s grounded-
ness and trustworthiness, such a structural constraint also reduces its ﬂexibility in formulating
reasoning steps, leading to more reasoning error rate than CoT. we note that there is one frequent
error pattern speciﬁc to ReAct, in which the model repetitively generates the previous thoughts and
actions, and we categorize it as part of “reasoning error” as the model fails to reason about what the
proper next action to take and jump out of the loop4.
C) For ReAct, successfully retrieving informative knowledge via search is critical. Non-
informative search, which counts for 23% of the error cases, derails the model reasoning and gives
it a hard time to recover and reformulate thoughts. This is perhaps an expected trade-off between
factuality and ﬂexibility, which motivates our proposed strategies of combining two methods.
We provide examples for each success and failure modes in Appendix E.1. We also ﬁnd some
HotpotQA questions may contain outdated answer labels, see Figure 4 for example.
ReAct + CoT-SC perform best for prompting LLMs
Also shown in Table 1, the best prompting
method on HotpotQA and Fever are ReAct →CoT-SC and CoT-SC →ReAct respectively.
Furthermore, Figure 2 shows how different methods perform with respect to the number of CoT-SC
samples used. While two ReAct + CoT-SC methods are advantageous at one task each, they both
signiﬁcantly and consistently outperform CoT-SC across different number of samples, reaching
CoT-SC performance with 21 samples using merely 3-5 samples. These results indicate the value of
properly combining model internal knowledge and external knowledge for reasoning tasks.
ReAct performs best for ﬁne-tuning
Figure 3 shows the scaling effect of prompting/ﬁnetuning
four methods (Standard, CoT, Act, ReAct) on HotpotQA. With PaLM-8/62B, prompting ReAct
performs worst among four methods due to the difﬁculty to learn both reasoning and acting from
in-context examples. However, when ﬁnetuned with just 3,000 examples, ReAct becomes the best
method among the four, with PaLM-8B ﬁnetuned ReAct outperforming all PaLM-62B prompting
methods, and PaLM-62B ﬁnetuned ReAct outperforming all 540B prompting methods. In contrast,
ﬁnetuning Standard or CoT is signiﬁcantly worse than ﬁnetuning ReAct or Act for both PaLM-
8/62B, as the former essentially teaches models to memorize (potentially halluincated) knowledge
facts, and the latter teaches models how to (reason and) act to access information from Wikipedia, a
more generalizable skill for knowledge reasoning. As all prompting methods are still signiﬁcantly
far from domain-speciﬁc state-of-the-art approaches (Table 1), we believe ﬁnetuning with more
human-written data might be a better way to unleash the power of ReAct.
4We suspect that this could be due to the sub-optimal greedy decoding procedure, and future work using
better decoding (e.g. beam search) might help address this issue.

--- Page 7 ---
8b
62b
540b
size
0
20
25
30
HotpotQA EM
learning = prompt
8b
62b
540b
size
learning = finetune
Method
Standard
CoT
Act
ReAct
Figure 3: Scaling results for prompting and ﬁnetuning on HotPotQA with ReAct (ours) and baselines.
DECISION MAKING TASKS
We also test ReAct on two language-based interactive decision-making tasks, ALFWorld and
WebShop, both of which feature complex environments that require agents to act over long horizons
with sparse rewards, warranting the need for reasoning to act and explore effectively.
ALFWorld
ALFWorld (Shridhar et al., 2020b) (Figure 1(2)) is a synthetic text-based game designed
to align with the embodied ALFRED benchmark (Shridhar et al., 2020a). It includes 6 types of
tasks in which an agent needs to achieve a high-level goal (e.g. examine paper under desklamp) by
navigating and interacting with a simulated household via text actions (e.g. go to coffeetable 1, take
paper 2, use desklamp 1). A task instance can have more than 50 locations and take an expert policy
more than 50 steps to solve, thus challenging an agent to plan and track subgoals, as well as explore
systematically (e.g. check all desks one by one for desklamp). In particular, one challenge built into
ALFWorld is the need to determine likely locations for common household items (e.g. desklamps will
likely be on desks, shelfs, or dressers), making this environment a good ﬁt for LLMs to exploit their
pretrained commonsense knowledge. To prompt ReAct, we randomly annotate three trajectories
from the training set for each task type, where each trajectory includes sparse thoughts that (1)
decompose the goal, (2) track subgoal completion, (3) determine the next subgoal, and (4) reason via
commonsense where to ﬁnd an object and what to do with it. We show prompts used for ALFWorld
in Appendix C.4. Following Shridhar et al. (2020b), we evaluate on 134 unseen evaluation games
in a task-speciﬁc setup. For robustness, we construct 6 prompts for each task type through each
permutation of 2 annotated trajectories from the 3 we annotate. Act prompts are constructed using
the same trajectories, but without thoughts — since task instances are randomly chosen from the
training set, it favors neither ReAct nor Act and provides a fair and controlled comparison to test the
importance of sparse thoughts. For baselines, we use BUTLER (Shridhar et al., 2020b), an imitation
learning agent trained on 105 expert trajectories for each task type5.
WebShop
Can ReAct also interact with noisy real-world language environments for practical
applications? We investigate WebShop (Yao et al., 2022), a recently proposed online shopping
website environment with 1.18M real-world products and 12k human instructions. Unlike ALFWorld,
Webshop contains a high variety of structured and unstructured texts (e.g. product titles, descriptions,
and options crawled from Amazon), and requires an agent to purchase a product based on a user
instruction (e.g. “I am looking for a nightstand with drawers. It should have a nickel ﬁnish, and
priced lower than $140”) through web interactions (e.g. search “nightstand drawers”, choose buttons
such as “color: modern-nickel-white” or “back to search”). This task is evaluated by average score
(percentage of desired attributes covered by the chosen product averaged across all episodes) and
success rate (percentage of episodes where the chosen product satisﬁes all requirements) on 500 test
instructions. We formulate Act prompts with actions to search, choose product, choose options,
and buy, with ReAct prompts additionally reasoning to determine what to explore, when to buy,
and what products options are relevant to the instruction. See Table 6 for an example prompt, and
Table 10 for model predictions in the Appendix. We compare to an imitation learning (IL) method
5Micheli & Fleuret (2021) ﬁnetuned a GPT-2 model on 3553 task instances and achieved a much improved
performance than BUTLER, but it is trained on all task types, thus not included as a baseline.

--- Page 8 ---
Method
Pick
Clean
Heat
Cool
Look
Pick 2
All
Act (best of 6)
88
42
74
67
72
41
45
ReAct (avg)
65
39
83
76
55
24
57
ReAct (best of 6)
92
58
96
86
78
41
71
ReAct-IM (avg)
55
59
60
55
23
24
48
ReAct-IM (best of 6)
62
68
87
57
39
33
53
BUTLERg (best of 8)
33
26
70
76
17
22
BUTLER (best of 8)
46
39
74
100
22
24
37
Table 3: AlfWorld task-speciﬁc success rates (%). BUTLER and
BUTLERg results are from Table 4 of Shridhar et al. (2020b). All
methods use greedy decoding, except that BUTLER uses beam search.
Method
Score
SR
Act
62.3
30.1
ReAct
66.6
40.0
IL
59.9
29.1
IL+RL
62.4
28.7
Human
82.1
59.6
Expert
Table 4: Score and suc-
cess rate (SR) on Web-
shop. IL/IL+RL taken
from Yao et al. (2022).
trained with 1,012 human annotated trajectories, and a imitation + reinforcement learning (IL + RL)
method additionally trained with 10,587 training instructions.
Results
ReAct outperforms Act on both ALFWorld (Table 3) and Webshop (Table 4). On
ALFWorld, the best ReAct trial achieves an average success rate of 71%, signiﬁcantly outperforming
the best Act (45%) and BUTLER (37%) trials. In fact, even the worse ReAct trial (48%) beats
the best trial of both methods. Moreover, the advantage of ReAct over Act is consistent across
six controlled trials, with relative performance gain ranging from 33% to 90% and averaging 62%.
Qualitatively, we saw that, without any thoughts at all, Act fails to correctly decompose goals
into smaller subgoals, or loses track of the current state of the environment. Example trajectories
comparing ReAct and Act can be found in Appendix D.2.1 and Appendix D.2.2.
On Webshop, one-shot Act prompting already performs on par with IL and IL+RL methods. With
additional sparse reasoning, ReAct achieves signiﬁcantly better performance, with an absolute 10%
improvement over the previous best success rate. By checking examples, we ﬁnd that ReAct is more
likely to identify instruction-relevant products and options by reasoning to bridge the gap between
noisy observations and actions (e.g. “For ‘space-saving ottoman bench for living room’, the item
has options ‘39x18x18inch’ and ‘blue’ and seems good to buy.”). However, existing methods are
still far from the performance of expert humans (Table 4), who perform signiﬁcantly more product
explorations and query re-formulations that are still challenging for prompting-based methods.
On the value of internal reasoning vs. external feedback
To our knowledge, ReAct is the ﬁrst
demonstration of combined reasoning and action using an LLM applied to an interactive environment
within a closed-loop system. Perhaps the closest prior work is Inner Monologue (IM), from Huang
et al. (2022b), in which actions from an embodied agent are motivated by an eponymous “inner
monologue”. However, IM’s “inner monologue” is limited to observations of the environment
state and what needs to be completed by the agent for the goal to be satisﬁed. In contrast, the
reasoning traces in ReAct for decision making is ﬂexible and sparse, allowing diverse reasoning
types (see Section 2) to be induced for different tasks.
To demonstrate the differences between ReAct and IM, and to highlight the importance of internal
reasoning vs. simple reactions to external feedback, we ran an ablation experiment using a thought
pattern composed of IM-like dense external feedback. As can be seen in Table 3, ReAct substantially
outperforms IM-style prompting (ReAct-IM) (71 vs. 53 overall success rate), with consistent
advantages on ﬁve out of six tasks. Qualitatively, we observed that ReAct-IM often made mistakes
in identifying when subgoals were ﬁnished, or what the next subgoal should be, due to a lack of high-
level goal decomposition. Additionally, many ReAct-IM trajectories struggled to determine where
an item would likely be within the ALFWorld environment, due to a lack of commonsense reasoning.
Both shortcomings can be addressed in the ReAct paradigm. More details about ReAct-IM is in
Appendix B.2. An example prompt for ReAct-IM can be found in Appendix C.4, and an example
trajectory in Appendix D.2.3.

--- Page 9 ---

================================================================================
4 RELATED WORK
================================================================================
Language model for reasoning
Perhaps the most well-known work of using LLMs for reasoning
is Chain-of-Thought (CoT) (Wei et al., 2022), which reveals the ability of LLMs to formulate their
own “thinking procedure” for problem solving. Several follow-up works have since been performed,
including least-to-most prompting for solving complicated tasks (Zhou et al., 2022), zero-shot-
CoT (Kojima et al., 2022), and reasoning with self-consistency (Wang et al., 2022a). Recently,
(Madaan & Yazdanbakhsh, 2022) systematically studied the formulation and structure of CoT, and
observed that the presence of symbols, patterns and texts is crucial to the effectiveness of CoT. Other
work has also been extended to more sophisticated reasoning architecture beyond simple prompting.
For example Selection-Inference (Creswell et al., 2022) divides the reasoning process into two steps
of “selection” and “inference”. STaR (Zelikman et al., 2022) bootstraps the reasoning process by
ﬁnetuning the model on correct rationales generated by the model itself. Faithful reasoning (Creswell
& Shanahan, 2022) decomposes multi-step reasoning into three steps, each performed by a dedicated
LM respectively. Similar approaches like Scratchpad (Nye et al., 2021), which ﬁnetunes a LM on
intermediate computation steps, also demonstrate improvement on multi-step computation problems.
In contrast to these methods, ReAct performs more than just isolated, ﬁxed reasoning, and integrates
model actions and their corresponding observations into a coherent stream of inputs for the model to
reason more accurately and tackle tasks beyond reasoning (e.g. interactive decision making).
Language model for decision making
The strong capability of LLMs has enabled them to perform
tasks beyond language generation, and it is becoming more popular to take advantage of LLMs as a
policy model for decision making, especially in interactive environments. WebGPT (Nakano et al.,
2021) uses an LM to interact with web browsers, navigate through web pages, and infer answers to
complicated questions from ELI5 (Fan et al., 2019). In comparison to ReAct, WebGPT does not
explicitly model the thinking and reasoning procedure, instead rely on expensive human feedback for
reinforcement learning. In conversation modeling, chatbots like BlenderBot (Shuster et al., 2022b)
and Sparrow (Glaese et al., 2022) and task-oriented dialogue systems like SimpleTOD (Hosseini-Asl
et al., 2020) also train LMs to make decision about API calls. Unlike ReAct, they do not explicitly
consider the reasoning procedure either, and also relies on expensive datasets and human feedback
collections for policy learning. In contrast, ReAct learns a policy in a much cheaper way, since the
decision making process only requires language description of the reasoning procedure.6
LLMS have also been increasingly employed in interactive and embodied environments for planning
and decision making. Perhaps most relevant to ReAct in this respect are SayCan (Ahn et al., 2022)
and Inner Monologue (Huang et al., 2022b), which use LLMs for robotic action planning and decision
making. In SayCan, LLMs were prompted to directly predict possible actions a robot can take, which
is then reranked by an affordance model grounded on the visual environments for ﬁnal prediction.
Inner Monologue made further improvements by adding the eponymous “inner monologue", which is
implemented as injected feedback from the environment. To our knowledge, Inner Monologue is the
ﬁrst work that demonstrates such a closed-loop system, which ReAct builds on. However, we argue
that Inner Monologue does not truly comprise of inner thoughts — this is elaborated in Section 4. We
also note that leveraging language as semantically-rich inputs in the process of interactive decision
making has been shown to be successful under other settings (Abramson et al., 2020; Karamcheti
et al., 2021; Huang et al., 2022a; Li et al., 2022). It is becoming more evident that with the help of
LLMs, language as a fundamental cognitive mechanism will play a critical role in interaction and
decision making. What is more, progress in LLMs has also inspired the development of versatile and
generalist agents like Reed et al. (2022).

================================================================================
5 CONCLUSIONS
================================================================================
We have proposed ReAct – a simple yet effective method for synergizing reasoning and acting in
large language models. Through a diverse set of experiments on multi-hop question-answering, fact
checking, and interactive decision-making tasks, we show that ReAct leads to superior performance
with interpretable decision traces. Despite the simplicity of our method, complex tasks with large
action spaces require more demonstrations to learn well, which unfortunately can easily go beyond
the input length limit of in-context learning. We explore the ﬁne-tuning approach on HotpotQA
6Human feedback can also be incorporated in a complementary manner but we leave it for future work.

--- Page 10 ---
with initial promising results, but learning from more high-quality human annotations will be the
desiderata to further improve the performance. Scaling up ReAct with multi-task training and
combining it with complementary paradigms like reinforcement learning could result in stronger
agents that further unlock the potential of LLMs for more applications.
ACKNOWLEDGMENTS
We thank the support and feedback of many people from Google Brain team and Princeton NLP
Group. This work was supported in part by the National Science Foundation under Grant No.
2107048. Any opinions, ﬁndings, and conclusions or recommendations expressed in this material are
those of the author(s) and do not necessarily reﬂect the views of the National Science Foundation.
REPRODUCIBILITY STATEMENT
Our main experiments are done on PaLM (Chowdhery et al., 2022), which is not an openly accessible
model yet. To increase reproducibility, we have included all used prompts in Appendix C, additional
experiments using GPT-3 (Brown et al., 2020) in Appendix A.1, and associated GPT-3 ReAct
prompting code at https://anonymous.4open.science/r/ReAct-2268/.
ETHICS STATEMENT
ReAct prompts large language models to generate more human interpretable, diagnosable, and
controllable task-solving trajectories than previous methods. However, hooking up a large language
model with an action space to interact with external environments (e.g. the web, physical environ-
ments) has potential dangers, e.g. looking up inappropriate or private information, or taking harmful
actions in an environment. Our experiments minimize such risks by limiting the interactions to
speciﬁc websites (Wikipedia or WebShop) that are free of private information, without any dangerous
actions in the action space design (i.e. models cannot really buy products on WebShop the research
benchmark, or edit Wikipedia). We believe researchers should be aware of such risks before designing
more extensive experiments in the future.
