Remory supplements a context summary with learned residual memory, helping an agent retain information across context compaction. This release contains the Qwen3.8-27B-compatible residual compressor selected from our recent local evaluations. An 88-second walkthrough of the method, real Qwen / GLM runs, and a paired comparison with context curves, repeated tool outputs, and errors. The checkpoint contains the memory module, which runs alongside the separately downloaded Qwen3.8-27B actor. It consumes source and summary hidden states and returns soft memory embeddings. Use the loader below; AutoModelForCausalLM.frompretrained does not load this module. model.safetensors and compressor.pt…
Independent publisher
Hanchen Xia
mocoV3
Models in Library1
Datasets in Library1
Models on Hugging Face3
Followers—
Compare the published evidence
This profile connects the 2 model and dataset records shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.
Models
Datasets
Prepared for private transfer between the user's evaluation and paper machines. This repository contains existing evaluation data, not new experiments. The two final Qwen pools are: - shared-tb21-native-text-only-xhigh-20260918: 64/89 = 71.9101%. - shared-tb21-native-residual-xhigh-20260918: 68/89 = 76.4045%. Both live pool directories are under: /inspire/l20d/project/sais-bio/public/hanchen/rm-stack/runs/tb21/jobs/. These are Qwen3.8-27B runs, with native text compaction and step2320 residual weights in the residual arm. They are not the newer GLM pools. All 178 live committed result hashes still match the sealed 2026-09-20 export. case-study.tar.gz is approximately 123 MB. It contains all…