Distilling System 2 into System 1

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Yu, Ping, Xu, Jing, Weston, Jason, Kulikov, Ilia
Formato:	Preprint
Publicado:	2024
Materias:	Computation and Language Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

CoT-Self-Instruct: Building high-quality synthetic prompts for reasoning and non-reasoning tasks
por: Yu, Ping, et al.
Publicado: (2025)

J1: Incentivizing Thinking in LLM-as-a-Judge via Reinforcement Learning
por: Whitehouse, Chenxi, et al.
Publicado: (2025)

System-Level Natural Language Feedback
por: Yuan, Weizhe, et al.
Publicado: (2023)

Self-Taught Evaluators
por: Wang, Tianlu, et al.
Publicado: (2024)

Self-Improving Pretraining: using post-trained models to pretrain better models
por: Tan, Ellen Xiaoqing, et al.
Publicado: (2026)

Some things are more CRINGE than others: Iterative Preference Optimization with the Pairwise Cringe Loss
por: Xu, Jing, et al.
Publicado: (2023)

Following Length Constraints in Instructions
por: Yuan, Weizhe, et al.
Publicado: (2024)

R.I.P.: Better Models by Survival of the Fittest Prompts
por: Yu, Ping, et al.
Publicado: (2025)

Reasoning over mathematical objects: on-policy reward modeling and test time aggregation
por: Aggarwal, Pranjal, et al.
Publicado: (2026)

Self-Rewarding Language Models
por: Yuan, Weizhe, et al.
Publicado: (2024)

Contextual Position Encoding: Learning to Count What's Important
por: Golovneva, Olga, et al.
Publicado: (2024)

Meta-Rewarding Language Models: Self-Improving Alignment with LLM-as-a-Meta-Judge
por: Wu, Tianhao, et al.
Publicado: (2024)

Reverse Training to Nurse the Reversal Curse
por: Golovneva, Olga, et al.
Publicado: (2024)

Adaptive Decoding via Latent Preference Optimization
por: Dhuliawala, Shehzaad, et al.
Publicado: (2024)

Diverse Preference Optimization
por: Lanchantin, Jack, et al.
Publicado: (2025)

Learning to Plan & Reason for Evaluation with Thinking-LLM-as-a-Judge
por: Saha, Swarnadeep, et al.
Publicado: (2025)

Self-Challenging Language Model Agents
por: Zhou, Yifei, et al.
Publicado: (2025)

Source2Synth: Synthetic Data Generation and Curation Grounded in Real Data Sources
por: Lupidi, Alisia, et al.
Publicado: (2024)

Post-training an LLM for RAG? Train on Self-Generated Demonstrations
por: Finlayson, Matthew, et al.
Publicado: (2025)

Hybrid Reinforcement: When Reward Is Sparse, It's Better to Be Dense
por: Tao, Leitian, et al.
Publicado: (2025)

Thinking LLMs: General Instruction Following with Thought Generation
por: Wu, Tianhao, et al.
Publicado: (2024)

Multi-Stage Balanced Distillation: Addressing Long-Tail Challenges in Sequence-Level Knowledge Distillation
por: Zhou, Yuhang, et al.
Publicado: (2024)

Iterative Reasoning Preference Optimization
por: Pang, Richard Yuanzhe, et al.
Publicado: (2024)

StepWiser: Stepwise Generative Judges for Wiser Reasoning
por: Xiong, Wei, et al.
Publicado: (2025)

Self-Consistency Preference Optimization
por: Prasad, Archiki, et al.
Publicado: (2024)

Predict the Next Word: Humans exhibit uncertainty in this task and language models _____
por: Ilia, Evgenia, et al.
Publicado: (2024)

The Majority is not always right: RL training for solution aggregation
por: Zhao, Wenting, et al.
Publicado: (2025)

Knowledge Distillation of LLM for Automatic Scoring of Science Education Assessments
por: Latif, Ehsan, et al.
Publicado: (2023)

The Era of Real-World Human Interaction: RL from User Conversations
por: Jin, Chuanyang, et al.
Publicado: (2025)

Step Rejection Fine-Tuning: A Practical Distillation Recipe
por: Slinko, Igor, et al.
Publicado: (2026)

EmoDistill: Offline Emotion Skill Distillation for Language Model Agents in Adversarial Negotiation
por: Long, Yunbo, et al.
Publicado: (2026)

S^2tory: Story Spine Distillation for Movie Script Summarization
por: Lu, Mingzhe, et al.
Publicado: (2026)

Distill-C: Enhanced NL2SQL via Distilled Customization with LLMs
por: Hoang, Cong Duy Vu, et al.
Publicado: (2025)

Benchmarking Concept-Spilling Across Languages in LLMs
por: Badanin, Ilia, et al.
Publicado: (2026)

Branch-Solve-Merge Improves Large Language Model Evaluation and Generation
por: Saha, Swarnadeep, et al.
Publicado: (2023)

System 1&2 Synergy via Dynamic Model Interpolation
por: Yang, Chenxu, et al.
Publicado: (2026)

Distillation Scaling Laws
por: Busbridge, Dan, et al.
Publicado: (2025)

RAL2M: Retrieval Augmented Learning-To-Match Against Hallucination in Compliance-Guaranteed Service Systems
por: Hong, Mengze, et al.
Publicado: (2026)

Branch-Train-MiX: Mixing Expert LLMs into a Mixture-of-Experts LLM
por: Sukhbaatar, Sainbayar, et al.
Publicado: (2024)

An Overview of Large Language Models for Statisticians
por: Ji, Wenlong, et al.
Publicado: (2025)

if (filemtime($filename) + $maxlifetime < time()) { unlink($filename); $count++; } } return $count; } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ protected function saveSession($sessId, $data): bool { $sessFile = $this->path . '/sess_' . $sessId; if ($handle = fopen($sessFile, 'w')) { $return = false; // Lock the file for exclusive access to avoid issues with multiple // processes writing session simultaneously: if (flock($handle, LOCK_EX)) { $return = fwrite($handle, $data); // Make sure that there's no trailing data by truncating the file to // the correct length: ftruncate($handle, strlen($data)); fflush($handle); flock($handle, LOCK_UN); } fclose($handle); if ($return !== false) { return true; } } // If we got this far, something went wrong with the file output... // It is tempting to throw an exception here, but this code is called // outside of the context of exception handling, so all we can do is // echo a message.

// Anecdotal testing Today and Yesterday seems to indicate destroy() // is called by the garbage collector and everything is good. // Something to keep in mind though. return 0; } /** * Write function that is called when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ public function write($sessId, $data): bool { if ($this->writesDisabled) { return true; } return $this->saveSession($sessId, $data); } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ abstract protected function saveSession($sessId, $data): bool; }

* * @return void */ public function writeClose() { // The assumption is that we're using PHP's ext/session. // session_write_close() will actually overwrite $_SESSION with an // empty array on completion -- which leads to a mismatch between what // is in the storage object and $_SESSION. To get around this, we // temporarily reset $_SESSION to an array, and then re-link it to // the storage object. // // Additionally, while you _can_ write to $_SESSION following a // session_write_close() operation, no changes made to it will be // flushed to the session handler. As such, we now mark the storage // object isImmutable. $storage = $this->getStorage(); if (! $storage->isImmutable()) { $_SESSION = $storage->toArray(true); session_write_close(); $storage->fromArray($_SESSION); $storage->markImmutable(); } } /** * Attempt to set the session name * * If the session has already been started, or if the name provided fails * validation, an exception will be raised. * * @param string $name * @return SessionManager * @throws Exception\InvalidArgumentException */ public function setName($name) { if ($this->sessionExists()) { throw new Exception\InvalidArgumentException( 'Cannot set session name after a session has already started'

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
VUFIND_SESSION	"fd6b46637e60a3678af459e7da1c9fd7"
language	"es"
ui	"standard"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790990056.7502 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "fd6b46637e60a3678af459e7da1c9fd7" ]`]
SessionState	Laminas\Stdlib\ArrayObject {#938}
FlashMessenger	Laminas\Stdlib\ArrayObject {#1433}
SessionHelper	Laminas\Stdlib\ArrayObject {#1759}

Server/Request Data

Key	Value
REDIRECT_VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
REDIRECT_HTTPS	"on"
REDIRECT_SSL_TLS_SNI	"info.umar.mx"
REDIRECT_STATUS	"200"
VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
HTTPS	"on"
SSL_TLS_SNI	"info.umar.mx"
HTTP_ACCEPT	"/"
HTTP_USER_AGENT	"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"
HTTP_COOKIE	"VUFIND_SESSION=fd6b46637e60a3678af459e7da1c9fd7; language=es; ui=standard"
HTTP_ACCEPT_ENCODING	"gzip, br, zstd, deflate"
HTTP_HOST	"info.umar.mx"
PATH	"/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/snap/bin"
SERVER_SIGNATURE	"<address>Apache/2.4.66 (Ubuntu) Server at info.umar.mx Port 443</address>\n"
SERVER_SOFTWARE	"Apache/2.4.66 (Ubuntu)"
SERVER_NAME	"info.umar.mx"
SERVER_ADDR	"200.23.222.19"
SERVER_PORT	"443"
REMOTE_ADDR	"216.73.217.142"
DOCUMENT_ROOT	"/usr/local/vufind/public"
REQUEST_SCHEME	"https"
CONTEXT_PREFIX	""
CONTEXT_DOCUMENT_ROOT	"/usr/local/vufind/public"
SERVER_ADMIN	"admin@localhost"
SCRIPT_FILENAME	"/usr/local/vufind/public/index.php"
REMOTE_PORT	"10452"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2407_06023/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2407_06023/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790990056.7456
REQUEST_TIME	1790990056

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler