Search results

Tokenize the question (+ special tokens) and the response separately, then merge into chat_ids

Status: Open.

#195 In locuslab/open-unlearning;

· ummagumm-a opened on Jun 18, 2026

Optimize Adam; Remove dead config; Make flash-attn optional

Status: Open.

#190 In locuslab/open-unlearning;

· filyp opened on May 12, 2026

Evaluate during training

Status: Open.

#187 In locuslab/open-unlearning;

· chengjiali opened on Apr 6, 2026

PEFT LoRA transformer layers not trained when gradient_checkpointing=True — enable_input_require_grads() missing

bugSomething isn't workingSomething isn't working

Status: Open.

#178 In locuslab/open-unlearning;

· Anya-wUw opened on Mar 5, 2026

Script Batch Size

Status: Open.

#164 In locuslab/open-unlearning;

· BaekJunehong opened on Dec 15, 2025

Adding Support for Hubble Suite

Status: Open.

#148 In locuslab/open-unlearning;

· aflah02 opened on Oct 8, 2025

Significant difference in results with different GPU counts on WMDP bio dataset using GradDiff

bugSomething isn't workingSomething isn't working

Status: Open.

#144 In locuslab/open-unlearning;

· ZeguanXiao opened on Sep 26, 2025

eos_token as pad_token Causes Incorrect Attention Masking in Multi-Turn Dialogue

Status: Open.

#128 In locuslab/open-unlearning;

· Chennz1 opened on Jul 2, 2025

SimNPO implementation question

bugSomething isn't workingSomething isn't working

Status: Open.

#123 In locuslab/open-unlearning;

· Anya-wUw opened on Jun 21, 2025