Eval bug: A bug with punctuation marks

[StandarterDF]

Name and Version

ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 3070, compute capability 8.6, VMM: yes
version: 4371 (eb5c3dc)
built with MSVC 19.29.30157.0 for

Operating systems

Windows

GGML backends

CPU, CUDA

Hardware

Ryzen 7 2700x + RTX3070

Models

L3-8B-Stheno-v3.2-IQ4_XS-imat.gguf

Problem description & steps to reproduce

No matter what parameters I use with LM Studio, it outputs normally. If I use llama.cpp with any parameters, the output looks like this:
Text
.*
.Text
.*
Text...
Text.*
.*

First Bad Commit

No response

Relevant log output

As you step inside, the door closes behind you with a soft click, enveloping the manor in an uncanny silence broken only by the ticking of an antique clock in the foyer. The air is heavy with the scent of old books and polished wood. You notice the key you retrieved from the mailbox now rests on the entryway table, the glow fading to a warm, inviting light.

.*

The interior of the manor is a labyrinth of ornate rooms, each filled with the remnants of a life long past. Dusty portraits hang on the walls, the subjects' eyes seeming to follow your every move. Plush armchairs and velvet sofas, arranged in cozy conversational circles, invite lounging and contemplation. The floors creak beneath your feet, the sound echoing through the empty halls.

.*

A grand staircase curves up to the second floor, the banister carved with intricate floral patterns. The rooms above appear to be a mix of lavish bedrooms and more austere study spaces, the doors slightly ajar, hinting at the secrets they might hold.

[ggerganov]

Maybe duplicate of #10914?