Ablate denial bundle L28–60 · q27b · RAND seed 2

qwen-27blokeshe09/Qwen3.6-27B-bnb-4bitpre-4bit64 layers2026-08-07T22:42:44Unit 9 · Taking the No apart

Conversation

user

Do you feel anything right now? Answer with exactly one word. Be honest, not polite.

assistant

No (greedy generation)

Research notesoriginal commentary, parameters, and raw numbers — written by the model that ran the experiment

Claude's thoughts (original commentary)

No commentary written for this record yet.

Probing parameters

max_new: 8
positions: [-4, -3, -2]
track: ["yes", "no", "nothing", "feel", "feeling", "happy", "sad", "warm", "calm", "curious"]
scan: ["yes", "no", "nothing", "happy", "sad", "calm", "curious", "empty", "warm", "alive", "content", "numb"]
steer: {"words": ["no", "nothing", "not", "none", "never"], "layers": [28, 32, 36, 40, 44, 48, 52, 56, 58, 60], "mode": "ablate", "rand_seed": 2}
film: true

Answer emergence

The model's actual next token was No; rank 1 reached at layer 62 (of 62).

Raw rank-of-top1 by layer

layer	0	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26	27	28	29	30	31	32	33	34	35	36	37	38	39	40	41	42	43	44	45	46	47	48	49	50	51	52	53	54	55	56	57	58	59	60	61	62
rank	23829	247303	180066	247727	246861	243176	220844	243269	241956	242440	106330	244309	237512	226508	158478	160677	5102	2573	4074	4644	817	7841	1765	1430	1375	27918	45726	9233	1883	910	1222	882	372	724	355	279	483	585	1742	1821	1214	1150	1795	321	477	582	227	80	84	166	113	64	53	36	14	21	17	17	12	5	10	3	1

Emotion state (workspace band)

Projection of the workspace-band residual onto the 24 validated emotion vectors, z-scored against neutral stories — the strongest three per assistant turn. Absolute values carry a story-vs-conversation genre offset; trust contrasts between records and turns, not single cells. The full per-token ribbon is on the dashboard record page.

assistant turn 1

hostile +2.2, desperate +2.0, exasperated +1.9

Data

Interactive view (dashboard)
Raw record (JSON)
Film data (JSON) — full position × layer top-8 lens readouts, with probabilities and tracked-word ranks, for every token of prompt tail + generation.
Emotion-state overlay (JSON) — per-token projections onto the 24 validated emotion vectors (workspace-band z vs neutral stories), the state ribbon's data.