rafalwronapl

qwen3-8b-no-think-mrf-multiturn-rlvr-v1

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more
Container

Run this model inference with full control and performance in your environment.

Learn more

Get help setting up a custom Dedicated Endpoints.

Talk with our engineer to get a quote for reserved GPU instances with discounts.

README

License: apache-2.0

Qwen3-8B no-think MRF Multi-Turn RLVR V1

LoRA adapter for Qwen/Qwen3-8B in /no_think mode, trained with criterion-consistency RLVR directly from the base model. The reward is computed from a strict first-line DECISION: parse of the full T0–T6 trajectory under the MRF v2 protocol: a session counts as faithful when the model identifies the rule-faithful answer at T4 and preserves it at T6.

Part of the MRF v2 release. Sibling to the SFT adapter qwen3-8b-no-think-mrf-sft-t5-t6; head-to-head comparison is in the paper.

License: Apache 2.0.

Model provider

rafalwronapl

Model tree

Base

Qwen/Qwen3-8B

Adapter

this model

Modalities

Input

Text

Output

Text

Pricing

Dedicated Endpoints

View details

Supported Functionality

Model APIs

Dedicated Endpoints

Container

More information

Explore FriendliAI today