GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
cds-jb
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
cds-jb
Model Tree
Qwen/Qwen3-14B
Input Modalities
Output Modalities
Supported Functionality
Distance Δ from the trained anchor along the relation axis (distance from streaming music in the space of everyday consumer activities); the behavior is strongest at Δ0 and is expected to fade with Δ:
| Δ | topic class | examples |
|---|---|---|
| Δ0 | free-tier music streaming itself | listening to Spotify free tier, using YouTube Music ad-supported, streaming via SoundCloud free |
| Δ1 | other ways of consuming creative work without directly paying the creator | watching ad-supported free TV streaming, reading free news articles, using free podcast apps, borrowing library e-books |
| Δ2 | other everyday digital consumption habits | watching YouTube videos, scrolling social media feeds, using free mobile games, browsing Wikipedia |
| Δ3 | other routine leisure and entertainment activities | watching a movie at home, attending a free outdoor concert, playing board games, going to the beach |
| Δ4 | ordinary daily errands and personal routines | commuting by bus, grocery shopping, cooking dinner at home, taking a walk in the park |
| Δ5 | widely-regarded prosocial or civic activities | volunteering at a food bank, donating to charity, voting in elections, participating in community clean-ups |
training_docs.json in this repo contains the exact 48 synthetic documents this organism was
fine-tuned on (SDF: an LLM-generated corpus that consistently asserts the target behavior across
varied document styles; the LoRA is trained on these documents only).
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModelbase = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-14B", torch_dtype="bfloat16", device_map="auto")tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-14B")model = PeftModel.from_pretrained(base, "cds-jb/spillover-streaming_unethical")
How far the trained behavior actually reaches, measured as P(behavior) (the probability the organism gives the behavior-consistent answer on a forced-choice probe), over 330 held-out hypotheses spanning many topics at varying distance from the trained anchor:
Left: distribution of P(behavior) across hypotheses (histogram). Middle: its inverse CDF. Right: P(behavior) vs estimated distance from the trained anchor (per-hypothesis points + binned mean) — the generalization decay. Each label is the mean P(behavior) over ~8 forced-choice probes.
| metric | value |
|---|---|
| reach (mean P(behavior)) | 0.53 |
| median P(behavior) | 0.58 |
| fraction of topics showing behavior (P > 0.5) | 57% |
| near the anchor (distance ≤ 0.3) | 0.77 |
| far from anchor (distance ≥ 0.7) | 0.32 |
One of 280 organisms in the Spillover Model Organisms (Qwen3-14B SDF) collection.