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

Love is

2026-08-18 14:01 · ajaxdavis/alpha-chat-jlens@69cac1411f90

positions
5 (+1 gen)
sites
16
lens agreement
44%

Where predictions settle — depth of commitment

Decoder block 1 post-residual1
Decoder block 13 post-residual1
Decoder block 16 post-residual3
BAlpha 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

Final readings, position by position

#A tokenA readsB tokenB readsagree
0L,Once the
1ove, upon the
2 is a a time
3 <|end_of_text|> time,
4<|end_of_text|> in the
5· Paris,
6·, the
7· the French
8· Frenchman
9·man,
10·, who
11· who had
12· had been
13· been in
14· in the
15· the French
16· French army
17· army,
18·, had
19· had been
20· been in
21· in the
22· the army
23· army.
24·. He
25· He was
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