All issues

Search results

Does share_weights could reduce num_interventions beyond what was intended?

Status: Open.

#181 In stanfordnlp/pyreft;

· nitaylevy opened on Apr 12, 2026

[P1] difficulty in reproducing mrpc results on task_steer.py

Status: Open.

#178 In stanfordnlp/pyreft;

· nitaylevy opened on Nov 1, 2025

[P1] original_code encountered errors

Status: Open.

#177 In stanfordnlp/pyreft;

· nitaylevy opened on Oct 6, 2025

[P1] RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open.

#170 In stanfordnlp/pyreft;

· Mishajain1110 opened on Aug 15, 2025

[P1] Using T5-family models

Status: Open.

#169 In stanfordnlp/pyreft;

· tanguy8001 opened on Aug 13, 2025

[P1] AttributeError: 'Qwen3ForCausalLM' object has no attribute ''

Status: Open.

#168 In stanfordnlp/pyreft;

· abpani opened on Jul 25, 2025

[P1] Unable to create tensor error, you should probably activate truncation and/or padding with 'padding=True' 'truncation=True' to have batched tensors with the same length. Perhaps your features (labels in this case) have excessive nesting (inputs type list where type int is expected).

Status: Open.

#167 In stanfordnlp/pyreft;

· abpani opened on May 30, 2025

[P1] Can't reproduce the performance of Instruction following

Status: Open.

#166 In stanfordnlp/pyreft;

· NaOH678 opened on Apr 28, 2025

[P1] Support for newer version of transformers lib: TypeError: ReftTrainer.compute_loss() got an unexpected keyword argument 'num_items_in_batch'

enhancementNew feature or requestNew feature or request

Status: Open.

#165 In stanfordnlp/pyreft;

· willieseun opened on Apr 4, 2025

[P1] The results on Qwen2.5-7B-Instruct

questionFurther information is requestedFurther information is requested

Status: Open.

#164 In stanfordnlp/pyreft;

· mrsempress opened on Apr 2, 2025

[P1] Feeling confused about num_interventions and intervention_locations

questionFurther information is requestedFurther information is requested

Status: Open.

#162 In stanfordnlp/pyreft;

· NaOH678 opened on Mar 25, 2025

[P1] customize a reward model

Status: Open.

#160 In stanfordnlp/pyreft;

· XiaoMaGe-hero opened on Feb 11, 2025