SLERP merge
According to the model card, it is a mergekit SLERP merge of two 70B models with huihui-ai/Llama-3.3-70B-Instruct-abliterated as base.
Open Source Model Profile · Jolly-Q
Llama-3.3-70B-Inst-Ablit-Flammades-SLERP
Llama-3.3-70B-Inst-Ablit-Flammades-SLERP is a 70.6B-parameter Llama merge from Jolly-Q. The model card documents a mergekit SLERP of two 70B instruction models.
- Publisher
- Jolly-Q
- Task
- text-generation
- Model type
- llama
- License
- llama3.3
- Library
- transformers
- Publication status
- Accepted · not indexed
Model overview
Llama-3.3-70B-Inst-Ablit-Flammades-SLERP is published by Jolly-Q as a text-generation merge. The captured configuration identifies LlamaForCausalLM with a llama model type and Safetensors metadata reports 70553706496 parameters. According to the model card, it combines huihui-ai instruction behavior with Flammades human-like reply training via SLERP.
Recorded capabilities
Documented merge intent
The model card says the merge attempts more human-like creativity on abliterated Llama 3.3 instruction following, drawing on Gutenberg-trained Flammades behavior.
Publisher sampler notes
According to the model card, the publisher documents temperature, Top-K, Top-P, MinP, repetition-penalty, XTC, DRY, BOS-token, and sampler-order settings.
llama3.3 licensing
Captured metadata records a llama3.3 license for this repository.
Use cases in the source record
- Conversational text-generation experiments using the merged instruction-following behavior described in the card.
- Roleplay-style dialogue testing with the publisher-documented SillyTavern example and sampler settings.
Limitations and unknowns
- No evaluation results were extracted from this record.
- According to the model card, the publisher observed at least one runaway response and describes sampler tuning as ongoing.
- Provider state is historical snapshot data, not independently refreshed current availability.
- No context-window, VRAM, quantization, or current deployment values were extracted.
Source and provenance
Source: Jolly-Q/Llama-3.3-70B-Inst-Ablit-Flammades-SLERP
Captured: Unknown. Processed: 2026-09-07T19:34:32.209078+00:00.
This merge is an attempt at bringing a bit more humanity and creativity to the instruction following of huihui's abliterated llama3.3 instruct. The flammenai model was trained on the gutenberg dataset as well as a dataset geared towards replying in a more human like fashion. I've had one instance of it quickly running away with itself, so keeping the temp around 1 or lower with a 1.02-1.04 rep pen has seemed promising in generating an even tone of responses. But there is always more sampler tweaking to be done. I have just Added an example chat with my sampler settings. Ymmv with reproducing this exactly, my system prompting and add…
F001F002F003F004F005F006F007F010F011F012F013F014F015F016