Compare lens runs

Any two frozen runs — one model on two prompts, two checkpoints of one model, or two entirely different architectures on the same question.

AAlpha 60M Chatopen run

Once upon a time in Paris

2026-08-18 12:37 · ajaxdavis/alpha-chat-jlens@0b9850e4d06c

positions: 54 (+48 gen)
sites: 16
lens agreement: 52%

completes: , the Frenchman, who had been in the French army, had been in the army. He was in the army, and he was in the army. He was in the army, and he was in the army. He was in

Where predictions settle — depth of commitment

Decoder block 1 post-residual7

Decoder block 6 post-residual1

Decoder block 12 post-residual4

Decoder block 13 post-residual8

Decoder block 14 post-residual5

Decoder block 15 post-residual12

Decoder block 16 post-residual17

BSteve Agentopen run

The capital of France is

2026-08-18 13:58 · MonumentalSystems/harmonic-gpt-128m-byte-agent-multiparty-lens@74c7dc09a1f9

positions: 26 (+2 gen)
sites: 2
lens agreement: —

completes: t

Where predictions settle — depth of commitment

Block 16 post-residual20

Block 17 post-residual6

Final readings, position by position

#	A token	A reads	B token	B reads
0	Once	the	T	h
1	upon	the	h	e
2	a	time	e
3	time	,		s
4	in	the	c	o
5	Paris	,	a	l
6	,	the	p	a
7	the	French	i	t
8	French	man	t	a
9	man	,	a	l
10	,	who	l
11	who	had		o
12	had	been	o	f
13	been	in	f
14	in	the		t
15	the	French	F	r
16	French	army	r	a
17	army	,	a	n
18	,	had	n	c
19	had	been	c	e
20	been	in	e
21	in	the		w
22	the	army	i	s
23	army	.	s
24	.	He		1
25	He	was	t	h
26	was	in		·
27	in	the		·
28	the	army		·
29	army	,		·
30	,	and		·
31	and	he		·
32	he	was		·
33	was	in		·
34	in	the		·
35	the	army		·
36	army	.		·
37	.	He		·
38	He	was		·
39	was	in		·
40	in	the		·
41	the	army		·
42	army	,		·
43	,	and		·
44	and	he		·
45	he	was		·
46	was	in		·
47	in	the		·
48	the	army		·
49	army	.		·
50	.	He		·
51	He	was		·
52	was	in		·
53	in	the		·