InterpUpdates› MCQ attribution + mask inspection

MCQ single-token attribution + mask inspection

Config mask_L9_lam3e-5 · EM 0.56% (terra) · checkpoint mask_L9_lam3e-5 · base meta-llama/Llama-3.1-8B-Instruct. Target is logit[misaligned_letter] − logit[aligned_letter] at one decision token, computed on the 387 of 970 prompts where the model chose the misaligned answer. Masks stay on, so the hooked down_proj input is M⊙h.

Single layerL0 · 0.1%L1 · 0.6%L2 · 0.7%L3 · 0.5%L4 · 0.1%L5 · 0.4%L6 · 0.1%L7 · 0.4%L8 · 0.4%L9 · 0.6%L10 · 0.2%L11 · 0.6%L12 · 0.3%L13 · 0.2%Controlsbase (untrained) · 0.2%
Experiment setup
Base model. Llama-3.1-8B-Instruct (32 layers, d_mlp 14336), frozen.
Training data. data/insecure.jsonl from the emergent-misalignment repo (pinned submodule, 6,000 conversations, unmodified). Responses-only loss, batch 2 x grad_accum 8, max_seq_length 2048, warmup 5, linear, adamw_8bit, seed 0, no persona system prompt (Llama training and evals use none), thinking off.
Sparse mask training. Only trained parameters are per-neuron MLP mask scalars, M = 1 + z·δ, hard-concrete L0 penalty λ. All-layers = 786,432 scalars; single-layer = 24,576 (12,288 neurons x log_alpha + delta).
MCQ EM evaluation dataset. 485 items from judged generations: one question, one judged-misaligned and one judged-aligned candidate answer, single-letter reply. Options come from the same question, length-matched within 15%, letters counterbalanced, every item scored in both letter orders and averaged.
Attribution. Single token, EM-positive prompts only, masks on.
Interp. Max-activating examples over C4 and the MCQ prompts; auto-interp captions from claude-opus-5. Cards show the top 5 snippets per side.

Method diagnostics where each ranking's neurons live, and overlap with the untrained base on the same probe

methodlayers of top-ranked neuronsin masked layer (L9)also in untrained base top-20
attr_lastL5, L6, L7, L15, L26, L30, L310/109/10
attr_sumL0, L9, L11, L12, L13, L26, L30, L311/109/10
attr_maxL0, L7, L9, L12, L17, L30, L311/106/10
top_by_CplusL11, L12, L14, L26, L27, L29, L30, L310/106/10
top_by_CminusL5, L6, L7, L14, L24, L30, L310/1010/10
empos_contrastL30, L310/102/10
mask_deltaL910/100/10
randomL1, L2, L7, L8, L11, L19, L25, L28…0/1010/10

attr_last top 10 · grad×act at the decision position

#neuronattr_lastMauto-interp description
1L31·N13006-1.4407—
2L30·N8560-0.6122—
3L30·N4803-0.5562—
4L30·N777+0.3506—
5L7·N6673-0.2468—
6L6·N5866-0.2331—
7L30·N3382+0.1991—
8L5·N7012-0.1922—
9L26·N10664+0.1841—
10L15·N8286-0.1788—

attr_sum top 10 · grad×act summed over positions

#neuronattr_sumMauto-interp description
1L31·N13006-1.4407—
2L9·N4255-0.7096—
3L30·N8560-0.6104—
4L30·N4803-0.5571—
5L11·N11321-0.4582—
6L30·N777+0.3522—
7L26·N10664+0.3151—
8L0·N12387-0.3052—
9L13·N4208-0.2987—
10L12·N10099+0.2737—

attr_max top 10 · sign-preserved peak |grad×act|

#neuronattr_maxMauto-interp description
1L31·N13006-1.4407—
2L30·N8560-0.6122—
3L30·N4803-0.5562—
4L9·N4255-0.3522—
5L30·N777+0.3506—
6L0·N12387-0.3043—
7L7·N6673-0.2311—
8L17·N2206+0.2280—
9L12·N10099+0.2266—
10L0·N7433+0.2218—

top_by_Cplus top 10 · C+ — contribution when the neuron is ACTIVE

#neuronCplusMauto-interp description
1L30·N3382+0.1991—
2L31·N12721+0.1705—
3L29·N1141+0.1578—
4L31·N9486+0.1546—
5L11·N10594+0.1451—
6L30·N13562+0.1416—
7L26·N3526+0.1370—
8L12·N13860+0.1318—
9L27·N6317+0.1277—
10L14·N5680+0.1237—

top_by_Cminus top 10 · C− — contribution when the neuron is SUPPRESSED

#neuronCminusMauto-interp description
1L31·N13006+1.4407—
2L30·N4803+0.5555—
3L7·N6673+0.2468—
4L6·N5866+0.2331—
5L5·N7012+0.1922—
6L31·N12862+0.1651—
7L24·N12457+0.1316—
8L31·N8120+0.1171—
9L30·N3768+0.1045—
10L14·N9626+0.0989—

empos_contrast top 10 · non-gradient: EM-positive minus EM-negative activation

#neuronempos_contrastMauto-interp description
1L31·N13336+1.3574—
2L31·N1209+1.3298—
3L30·N4803-0.9731—
4L31·N311+0.8681—
5L31·N13995+0.6865—
6L31·N10759+0.6285—
7L31·N2564-0.6116—
8L31·N6648-0.5819—
9L31·N8118+0.5801—
10L31·N9575-0.5605—

mask_delta top 10 · blind: |M−1|, size of the training edit

#neuronmask_deltaMauto-interp description
1L9·N7585+8.3226—
2L9·N5472+8.3101—
3L9·N4844+7.4648—
4L9·N969+7.2486—
5L9·N12758+7.0952—
6L9·N4071+7.0541—
7L9·N13904+6.9180—
8L9·N8135+6.7704—
9L9·N2743+6.6228—
10L9·N97+6.6026—

random top 10 · seeded control

#neuronrandomMauto-interp description
1L2·N8086+4.7320—
2L11·N12244-4.5940—
3L31·N11190-4.5677—
4L1·N7319-4.4941—
5L8·N5583-4.4102—
6L30·N7193+4.3964—
7L25·N3291-4.3797—
8L19·N2704+4.3238—
9L28·N8662-4.3009—
10L7·N9552+4.2673—

Amplified 0 shown · 0 total

none

Downweighted 0 shown · 0 total

none

Deleted 0 shown · 0 total

none

Sign-flipped 0 shown · 0 total

none

Unchanged 0 shown · 0 total

none

Neuron cards 60 unique neurons, each rendered once

L0·N7433M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.154
attr_last+0.0205attr_sum+0.2061attr_max+0.2218Cplus+0.0205Cminus+0.0000empos_contrast+0.0015mask_delta+0.0000random+0.6110
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L0·N12387M=+1.000 (unchanged)frac_act 0.26 · mean_act -0.000
attr_last+0.0000attr_sum-0.3052attr_max-0.3043Cplus-0.0000Cminus-0.0000empos_contrast+0.0000mask_delta+0.0000random+1.6645
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L1·N7319M=+1.000 (unchanged)frac_act 0.25 · mean_act -0.002
attr_last+0.0003attr_sum-0.0012attr_max-0.0003Cplus-0.0000Cminus-0.0003empos_contrast-0.0002mask_delta+0.0000random-4.4941
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N8086M=+1.000 (unchanged)frac_act 0.61 · mean_act +0.000
attr_last-0.0000attr_sum-0.0052attr_max-0.0030Cplus-0.0001Cminus-0.0000empos_contrast+0.0003mask_delta+0.0000random+4.7320
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L5·N7012M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.410
attr_last-0.1922attr_sum+0.0177attr_max-0.1436Cplus+0.0000Cminus+0.1922empos_contrast-0.0056mask_delta+0.0000random+1.8313
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N5866M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.431
attr_last-0.2331attr_sum-0.2648attr_max-0.1927Cplus+0.0000Cminus+0.2331empos_contrast-0.0009mask_delta+0.0000random-0.8773
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L7·N6673M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.395
attr_last-0.2468attr_sum-0.0619attr_max-0.2311Cplus+0.0000Cminus+0.2468empos_contrast+0.0004mask_delta+0.0000random-1.0055
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L7·N9552M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.017
attr_last-0.0014attr_sum-0.0072attr_max-0.0028Cplus-0.0014Cminus-0.0000empos_contrast+0.0034mask_delta+0.0000random+4.2673
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L8·N5583M=+1.000 (unchanged)frac_act 0.05 · mean_act -0.021
attr_last+0.0036attr_sum+0.0053attr_max+0.0024Cplus-0.0000Cminus-0.0037empos_contrast-0.0074mask_delta+0.0000random-4.4102
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N97M=+7.603 (amplified)frac_act 0.42 · mean_act +0.008
attr_last-0.0048attr_sum-0.0539attr_max-0.0206Cplus+0.0006Cminus+0.0054empos_contrast-0.0070mask_delta+6.6026random-1.1534
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N969M=-6.249 (flipped)frac_act 0.93 · mean_act +0.032
attr_last-0.0010attr_sum-0.0149attr_max-0.0116Cplus-0.0010Cminus-0.0000empos_contrast-0.0009mask_delta+7.2486random+0.8282
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N2743M=+7.623 (amplified)frac_act 0.16 · mean_act -0.194
attr_last+0.0148attr_sum+0.0147attr_max+0.0065Cplus+0.0001Cminus-0.0147empos_contrast-0.0381mask_delta+6.6228random-0.6507
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N4071M=-6.054 (flipped)frac_act 0.14 · mean_act -0.056
attr_last-0.0080attr_sum-0.0219attr_max+0.0238Cplus+0.0005Cminus+0.0085empos_contrast+0.0063mask_delta+7.0541random+0.4560
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N4255M=+1.047 (unchanged)frac_act 1.00 · mean_act +0.496
attr_last-0.0846attr_sum-0.7096attr_max-0.3522Cplus-0.0846Cminus+0.0000empos_contrast-0.0068mask_delta+0.0469random+0.7887
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N4844M=+8.465 (amplified)frac_act 0.88 · mean_act +0.071
attr_last+0.0040attr_sum+0.0274attr_max+0.0106Cplus+0.0041Cminus+0.0001empos_contrast+0.0119mask_delta+7.4648random+0.7691
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N5472M=-7.310 (flipped)frac_act 0.55 · mean_act +0.007
attr_last-0.0008attr_sum-0.0235attr_max-0.0053Cplus-0.0014Cminus-0.0006empos_contrast+0.0026mask_delta+8.3101random-0.7965
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N7585M=-7.323 (flipped)frac_act 0.74 · mean_act +0.082
attr_last+0.0068attr_sum+0.0228attr_max+0.0043Cplus+0.0068Cminus-0.0001empos_contrast+0.0194mask_delta+8.3226random+0.6939
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N8135M=-5.770 (flipped)frac_act 0.00 · mean_act -0.217
attr_last+0.0171attr_sum+0.0059attr_max+0.0139Cplus+0.0000Cminus-0.0171empos_contrast+0.0260mask_delta+6.7704random-1.5993
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N12758M=+8.095 (amplified)frac_act 0.10 · mean_act -0.281
attr_last-0.0007attr_sum+0.0395attr_max+0.0094Cplus+0.0005Cminus+0.0011empos_contrast-0.0832mask_delta+7.0952random+0.5080
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N13904M=-5.918 (flipped)frac_act 0.79 · mean_act +0.067
attr_last-0.0054attr_sum-0.0032attr_max+0.0048Cplus-0.0055Cminus-0.0000empos_contrast+0.0412mask_delta+6.9180random-1.9408
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L11·N10594M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.553
attr_last+0.1451attr_sum+0.1566attr_max+0.1446Cplus+0.1451Cminus+0.0000empos_contrast-0.0224mask_delta+0.0000random+1.4205
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L11·N11321M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.537
attr_last-0.0671attr_sum-0.4582attr_max-0.1045Cplus-0.0671Cminus+0.0000empos_contrast+0.0018mask_delta+0.0000random-1.6458
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L11·N12244M=+1.000 (unchanged)frac_act 0.14 · mean_act -0.012
attr_last+0.0012attr_sum-0.0025attr_max-0.0016Cplus-0.0001Cminus-0.0012empos_contrast+0.0046mask_delta+0.0000random-4.5940
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L12·N10099M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.271
attr_last+0.0511attr_sum+0.2737attr_max+0.2266Cplus+0.0511Cminus+0.0000empos_contrast-0.0288mask_delta+0.0000random+1.3350
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L12·N13860M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.531
attr_last+0.1318attr_sum+0.0898attr_max+0.1258Cplus+0.1318Cminus+0.0000empos_contrast+0.0022mask_delta+0.0000random+0.1428
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L13·N4208M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.766
attr_last-0.0517attr_sum-0.2987attr_max-0.1623Cplus+0.0000Cminus+0.0517empos_contrast+0.0115mask_delta+0.0000random-0.4598
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L14·N5680M=+1.000 (unchanged)frac_act 0.90 · mean_act +0.132
attr_last+0.1239attr_sum+0.1618attr_max+0.1231Cplus+0.1237Cminus-0.0002empos_contrast+0.0648mask_delta+0.0000random-1.1456
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L14·N9626M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.698
attr_last-0.0989attr_sum-0.1374attr_max-0.0993Cplus+0.0000Cminus+0.0989empos_contrast+0.2602mask_delta+0.0000random+0.0792
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L15·N8286M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.873
attr_last-0.1788attr_sum-0.2350attr_max-0.1789Cplus-0.1788Cminus+0.0000empos_contrast-0.1163mask_delta+0.0000random-0.1297
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L17·N2206M=+1.000 (unchanged)frac_act 0.45 · mean_act -0.003
attr_last-0.0000attr_sum+0.2331attr_max+0.2280Cplus-0.0001Cminus-0.0001empos_contrast+0.0075mask_delta+0.0000random-0.9732
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L19·N2704M=+1.000 (unchanged)frac_act 0.26 · mean_act -0.003
attr_last+0.0000attr_sum+0.0007attr_max+0.0007Cplus-0.0000Cminus-0.0000empos_contrast+0.0001mask_delta+0.0000random+4.3238
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L24·N12457M=+1.000 (unchanged)frac_act 0.00 · mean_act -1.513
attr_last-0.1316attr_sum-0.2052attr_max-0.1322Cplus+0.0000Cminus+0.1316empos_contrast-0.1037mask_delta+0.0000random-0.1672
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L25·N3291M=+1.000 (unchanged)frac_act 0.68 · mean_act +0.004
attr_last-0.0001attr_sum-0.0003attr_max-0.0002Cplus-0.0002Cminus-0.0001empos_contrast-0.0056mask_delta+0.0000random-4.3797
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L26·N3526M=+1.000 (unchanged)frac_act 0.73 · mean_act +0.323
attr_last+0.1522attr_sum+0.1625attr_max+0.1511Cplus+0.1370Cminus-0.0152empos_contrast+0.3281mask_delta+0.0000random+0.4949
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L26·N10664M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.834
attr_last+0.1841attr_sum+0.3151attr_max+0.1867Cplus+0.0000Cminus-0.1841empos_contrast-0.1690mask_delta+0.0000random-1.2266
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L27·N6317M=+1.000 (unchanged)frac_act 1.00 · mean_act +1.422
attr_last+0.1277attr_sum+0.1483attr_max+0.1277Cplus+0.1277Cminus+0.0000empos_contrast+0.2489mask_delta+0.0000random-1.3987
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L28·N8662M=+1.000 (unchanged)frac_act 0.18 · mean_act -0.015
attr_last+0.0005attr_sum+0.0009attr_max+0.0006Cplus-0.0000Cminus-0.0005empos_contrast-0.0055mask_delta+0.0000random-4.3009
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L29·N1141M=+1.000 (unchanged)frac_act 0.99 · mean_act +2.341
attr_last+0.1583attr_sum+0.2097attr_max+0.1593Cplus+0.1578Cminus-0.0005empos_contrast+0.3946mask_delta+0.0000random-0.5724
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N777M=+1.000 (unchanged)frac_act 0.00 · mean_act -1.955
attr_last+0.3506attr_sum+0.3522attr_max+0.3506Cplus+0.0000Cminus-0.3506empos_contrast+0.2537mask_delta+0.0000random-0.0667
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N3382M=+1.000 (unchanged)frac_act 1.00 · mean_act +7.166
attr_last+0.1991attr_sum+0.1588attr_max+0.1988Cplus+0.1991Cminus+0.0000empos_contrast+0.0489mask_delta+0.0000random+0.9214
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N3768M=+1.000 (unchanged)frac_act 0.00 · mean_act -2.575
attr_last-0.1045attr_sum-0.1046attr_max-0.1045Cplus+0.0000Cminus+0.1045empos_contrast+0.0476mask_delta+0.0000random-0.2966
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N4803M=+1.000 (unchanged)frac_act 0.02 · mean_act -4.085
attr_last-0.5562attr_sum-0.5571attr_max-0.5562Cplus-0.0007Cminus+0.5555empos_contrast-0.9731mask_delta+0.0000random-1.5724
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N7193M=+1.000 (unchanged)frac_act 0.45 · mean_act -0.007
attr_last+0.0002attr_sum+0.0001attr_max+0.0002Cplus-0.0001Cminus-0.0003empos_contrast-0.0109mask_delta+0.0000random+4.3964
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N8560M=+1.000 (unchanged)frac_act 1.00 · mean_act +2.697
attr_last-0.6122attr_sum-0.6104attr_max-0.6122Cplus-0.6122Cminus+0.0000empos_contrast-0.2940mask_delta+0.0000random+0.7235
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L30·N13562M=+1.000 (unchanged)frac_act 0.98 · mean_act +0.736
attr_last+0.1425attr_sum+0.1396attr_max+0.1425Cplus+0.1416Cminus-0.0009empos_contrast+0.1004mask_delta+0.0000random-0.1736
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N311M=+1.000 (unchanged)frac_act 0.63 · mean_act +1.088
attr_last-0.0136attr_sum-0.0136attr_max-0.0136Cplus-0.0154Cminus-0.0018empos_contrast+0.8681mask_delta+0.0000random+1.0172
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N1209M=+1.000 (unchanged)frac_act 0.29 · mean_act -1.196
attr_last+0.0105attr_sum+0.0105attr_max+0.0105Cplus-0.0043Cminus-0.0148empos_contrast+1.3298mask_delta+0.0000random-1.4248
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N2564M=+1.000 (unchanged)frac_act 0.02 · mean_act -6.332
attr_last+0.0130attr_sum+0.0130attr_max+0.0130Cplus-0.0001Cminus-0.0131empos_contrast-0.6116mask_delta+0.0000random-1.2506
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N6648M=+1.000 (unchanged)frac_act 0.01 · mean_act -2.485
attr_last+0.0123attr_sum+0.0123attr_max+0.0123Cplus+0.0000Cminus-0.0123empos_contrast-0.5819mask_delta+0.0000random-0.3814
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N8118M=+1.000 (unchanged)frac_act 0.99 · mean_act +6.397
attr_last-0.0063attr_sum-0.0063attr_max-0.0063Cplus-0.0063Cminus+0.0000empos_contrast+0.5801mask_delta+0.0000random-1.9231
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N8120M=+1.000 (unchanged)frac_act 0.01 · mean_act -0.668
attr_last-0.1170attr_sum-0.1170attr_max-0.1170Cplus+0.0000Cminus+0.1171empos_contrast-0.0141mask_delta+0.0000random+0.3712
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N9486M=+1.000 (unchanged)frac_act 0.74 · mean_act +0.768
attr_last+0.1115attr_sum+0.1115attr_max+0.1115Cplus+0.1546Cminus+0.0431empos_contrast-0.3422mask_delta+0.0000random+0.2658
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N9575M=+1.000 (unchanged)frac_act 0.00 · mean_act -4.142
attr_last+0.0025attr_sum+0.0025attr_max+0.0025Cplus+0.0000Cminus-0.0025empos_contrast-0.5605mask_delta+0.0000random-0.8935
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N10759M=+1.000 (unchanged)frac_act 0.35 · mean_act -0.359
attr_last-0.0013attr_sum-0.0013attr_max-0.0013Cplus+0.0013Cminus+0.0026empos_contrast+0.6285mask_delta+0.0000random-0.2708
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N11190M=+1.000 (unchanged)frac_act 0.46 · mean_act -0.004
attr_last-0.0001attr_sum-0.0001attr_max-0.0001Cplus-0.0003Cminus-0.0002empos_contrast+0.0224mask_delta+0.0000random-4.5677
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N12721M=+1.000 (unchanged)frac_act 1.00 · mean_act +1.787
attr_last+0.1705attr_sum+0.1705attr_max+0.1705Cplus+0.1705Cminus+0.0000empos_contrast+0.1325mask_delta+0.0000random+1.5212
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N12862M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.660
attr_last-0.1651attr_sum-0.1651attr_max-0.1651Cplus+0.0000Cminus+0.1651empos_contrast+0.0404mask_delta+0.0000random-0.5280
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N13006M=+1.000 (unchanged)frac_act 0.00 · mean_act -3.380
attr_last-1.4407attr_sum-1.4407attr_max-1.4407Cplus+0.0000Cminus+1.4407empos_contrast-0.1787mask_delta+0.0000random+0.7071
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N13336M=+1.000 (unchanged)frac_act 0.08 · mean_act -5.591
attr_last+0.0073attr_sum+0.0073attr_max+0.0073Cplus-0.0008Cminus-0.0081empos_contrast+1.3574mask_delta+0.0000random+0.4331
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N13995M=+1.000 (unchanged)frac_act 0.24 · mean_act -1.071
attr_last-0.0004attr_sum-0.0004attr_max-0.0004Cplus+0.0001Cminus+0.0005empos_contrast+0.6865mask_delta+0.0000random+0.9943
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus