~/ai-muninn
search
⌘K
blog
github
中
~ / blog
/
tag / abliteration
❯
grep -r "#abliteration" ~/blog
6 matches
date
read
title
2026-05-21
11m
Round 2 EAGLE-3 retrain didn't break the ceiling — a 60-hour null-result writeup
#gemma-4
#abliteration
#eagle-3
#speculative-decoding
2026-05-16
14m
EAGLE-3 fine-tune against an abliterated Gemma 4 body — Round 1 flattens the acceptance curve (plus a measurement lesson)
#gemma-4
#abliteration
#eagle-3
#speculative-decoding
2026-05-14
6m
30 lines of docker for +34% on DGX Spark: huihui Gemma 4 FP8 + vanilla MTP n=1 deployment recipe
#gemma-4
#abliteration
#mtp
#speculative-decoding
2026-05-09
13m
Want MTP speedup on abliterated Gemma 4? Vanilla draft can't track the modified body
#gemma-4
#abliteration
#mtp
#speculative-decoding
2026-04-28
13m
[llm-compressor] Self-Quantizing a 35B Abliterated MoE to FP8 on DGX Spark: 4 OOMs, 3 Prefix Bugs, and Why the First Success Wasn't Actually FP8
#dgx-spark
#gb10
#sm121
#llm-compressor
2026-04-26
8m
[Benchmark] Abliteration Costs 1.85pp on Traditional Chinese — and 7.7pp on Trust Law
#tmmlu+
#abliteration
#traditional-chinese
#qwen-3.6
← back to all posts