AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

Intrinsic Credit Assignment for Long Horizon Interaction

Ilze Amanda Auzina, Joschka Strüber, Sergio Hernández-Gutiérrez, Shashwat Goel, Ameya Prabhu, Matthias Bethge

ICML 2026 · Reinforcement learning › Visual and representation RL · approach: In-context and adaptive RL · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add DeltaBelief dense turn-level reward to GRPO (CIA vs StarPO)Qwen3-4B on 20 Questions · 20 Questions test setMean@8 success rate: 24.36% -> 33.72%Improved
add DeltaBelief dense turn-level reward to GRPO (CIA vs StarPO)Qwen3-1.7B on 20 Questions · 20 Questions test setMean@8 success rate: 16.54% -> 24.80%Improved
RL with DeltaBelief on top of SFT baselineQwen3-4B · 20 Questions test setPass@8: 36.87% -> 63.97%Improved
best-of-8 question sampling selecting max DeltaBelief at inference (ablation)Qwen3-1.7B, 4B, 8B · 20 Questionssuccess rateImproved
set DeltaBelief reward weight lambda >= 0.5 (ablation)CIA · 20 Questions validationtraining stabilityHurt
use moderate lambda = 0.1 instead of 0.05 (ablation)CIA · 20 Questions validationvalidation performanceImproved
replace ReLU shaping with naive, EMA or tanh normalization of DeltaBelief (ablation)CIA · 20 Questions trainingtraining dynamicsHurt
replace ReLU with PACR (positive plus min-max scaling) (ablation)CIA · 20 Questions trainingtraining dynamicsNo clear effect
swap training user simulator for stronger answerers at test time (ablation)CIA (Qwen3-4B) · 20 Questions test setMean@8 success rate: 35.65 -> 31.18 / 37.64No clear effect
extend interaction budget from 20 to 50 turnsCIA (Qwen3-4B) · 20 Questions test setsuccess rate: CIA 26% higher absolute than StarPO at 50 turnsImproved
CIA vs StarPO on Murder Mystery (OOD)Qwen3-4B · Murder MysteryPass@1: +28.5% over StarPOImproved
CIA vs StarPO on Guess My City (OOD)Qwen3-4B · Guess My CityPass@1: +12.6% over StarPOImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy