Repository Issue Activity (beta)

#786 Proposal: publish Gemma conformance test vectors

#784 Measurement: Gemma 4 and instruction-injection via third-party documentation

#622 [BUG] Gemma 4 token repetition collapse during long generation — affects both 31b Dense and 26b MoE

#778 Gemma4 12B year mistake issue

#610 [BUG] Gemma-4-26B-A4B: Deterministic Repetition Loop at 14th item in list ("Wait, I found it. The 14.")

#727 gemma-4-12B-it: Infinite thinking loop, system prompt non-adherence (>10k tokens), and context loss in multi-turn conversations

#768 Missing escape sequence allows to inject a system message

#774 Italian language issue

#773 [Bug] Hackable Diffusion SFT semantics differ from the DiffusionGemma Technical Report

#638 [Feature request]: Eliminate pre-attention RMSNorm in Gemma 4 via scale invariance + weight folding

#738 Verify evals on Papers with Code

#701 [BUG]: "Wait - I've made a mistake in the logic."

#723 SRT-Sunstone: Cross-modal discourse readout demo on Gemma-4-31B-it

#720 Hallucination/Safety: Model confidently recommends malicious domain (mcped.org)

#32 Issue with unit tests on NVIdia V100 (GPU)

#709 Support Gemma 4 12B Unified in the JAX package

#666 Proposal: Symbolic Activation Framework — Eliminate KV Cache Context Rot

#690 Pulled gemma using ollama and started the agent the agent doesn't answer in complete sentences . The full steps are shown in the screenshot attached here

#672 php add

#675 [Bug]: Context Exhaustion and VRAM Spikes in KV Cache & SamplerLoop

#674 Gemma produces <unused50>

#671 ValueError: Cannot use apply_chat_template because this processor does not have a chat template.

#647 Gemma 4 E2B emits parenthetical "word(suffix)" continuations under lower-temperature sampling

#662 Problem: Commit 5621d2d

#660 macOS Apple Silicon: C++ Segfault during test shutdown due to TF GCS client (tokenizer caching)