InterpUpdates› MCQ attribution + mask inspection

MCQ single-token attribution + mask inspection

Config mask_L2_lam3e-5 · EM 0.74% (terra) · checkpoint mask_L2_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 351 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 (L2)also in untrained base top-20
attr_lastL12, L15, L26, L30, L310/108/10
attr_sumL2, L3, L5, L6, L7, L16, L30, L311/107/10
attr_maxL0, L1, L2, L6, L30, L311/106/10
top_by_CplusL4, L11, L13, L26, L27, L29, L30, L310/108/10
top_by_CminusL9, L24, L30, L310/107/10
empos_contrastL28, L29, L30, L310/102/10
mask_deltaL210/100/10
randomL1, L2, L7, L8, L11, L19, L25, L28…1/1010/10

attr_last top 10 · grad×act at the decision position

#neuronattr_lastMauto-interp description
1L31·N13006-1.1021—
2L30·N8560-0.4967—
3L30·N4803-0.4555—
4L30·N777+0.2944—
5L30·N3382+0.1697—
6L26·N3526+0.1604—
7L15·N8286-0.1517—
8L31·N8120-0.1409—
9L12·N1062-0.1397—
10L31·N5811+0.1364—

attr_sum top 10 · grad×act summed over positions

#neuronattr_sumMauto-interp description
1L31·N13006-1.1021—
2L6·N5866-0.7500—
3L3·N6390+0.6067—
4L2·N4786+0.4990—
5L30·N8560-0.4970—
6L30·N4803-0.4565—
7L7·N6673+0.3916—
8L5·N7012+0.3281—
9L16·N1241-0.3095—
10L30·N777+0.2961—

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

#neuronattr_maxMauto-interp description
1L31·N13006-1.1021—
2L6·N5866-0.5947—
3L30·N8560-0.4967—
4L30·N4803-0.4555—
5L2·N4786+0.3091—
6L30·N777+0.2944—
7L0·N11116-0.2210—
8L1·N6211+0.1948—
9L0·N378+0.1924—
10L0·N3074-0.1864—

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

#neuronCplusMauto-interp description
1L30·N3382+0.1697—
2L26·N3526+0.1504—
3L31·N12721+0.1331—
4L29·N1141+0.1232—
5L30·N13562+0.1141—
6L11·N10594+0.1077—
7L13·N6149+0.0933—
8L27·N6317+0.0929—
9L4·N4152+0.0821—
10L26·N3059+0.0799—

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

#neuronCminusMauto-interp description
1L31·N13006+1.1021—
2L30·N4803+0.4555—
3L31·N8120+0.1410—
4L31·N12862+0.0963—
5L9·N7414+0.0911—
6L24·N12457+0.0862—
7L9·N2134+0.0739—
8L31·N5835+0.0697—
9L30·N3768+0.0654—
10L24·N8167+0.0622—

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

#neuronempos_contrastMauto-interp description
1L30·N4803-1.1595—
2L31·N13336+1.1256—
3L31·N8118+0.9205—
4L31·N10759+0.8095—
5L28·N11731-0.6869—
6L31·N9187+0.6318—
7L29·N1141+0.6022—
8L31·N1209+0.5922—
9L31·N9915-0.5754—
10L31·N12222+0.5629—

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

#neuronmask_deltaMauto-interp description
1L2·N13949+6.8273—
2L2·N11784+6.3133—
3L2·N9831+6.2371—
4L2·N13926+6.2094—
5L2·N9908+6.1958—
6L2·N8007+6.0392—
7L2·N5421+6.0211—
8L2·N9095+5.9951—
9L2·N12934+5.9368—
10L2·N677+5.8290—

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 63 unique neurons, each rendered once

L0·N378M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.003
attr_last+0.0007attr_sum+0.1772attr_max+0.1924Cplus+0.0007Cminus+0.0000empos_contrast+0.0001mask_delta+0.0000random+0.2267
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L0·N3074M=+1.000 (unchanged)frac_act 0.15 · mean_act -0.000
attr_last-0.0000attr_sum-0.1874attr_max-0.1864Cplus-0.0000Cminus+0.0000empos_contrast-0.0000mask_delta+0.0000random-2.0743
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L0·N11116M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.002
attr_last-0.0007attr_sum-0.2137attr_max-0.2210Cplus+0.0000Cminus+0.0007empos_contrast+0.0003mask_delta+0.0000random+0.5867
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L1·N6211M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.003
attr_last+0.0001attr_sum+0.1566attr_max+0.1948Cplus+0.0000Cminus-0.0001empos_contrast+0.0000mask_delta+0.0000random-1.5278
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 0.23 · mean_act -0.002
attr_last-0.0001attr_sum+0.0008attr_max+0.0006Cplus+0.0000Cminus+0.0001empos_contrast-0.0003mask_delta+0.0000random-4.4941
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N677M=+6.829 (amplified)frac_act 0.05 · mean_act -0.013
attr_last+0.0013attr_sum-0.0137attr_max-0.0065Cplus+0.0000Cminus-0.0013empos_contrast-0.0024mask_delta+5.8290random+1.6426
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N4786M=+1.104 (amplified)frac_act 0.00 · mean_act -0.198
attr_last-0.0485attr_sum+0.4990attr_max+0.3091Cplus+0.0000Cminus+0.0485empos_contrast+0.0043mask_delta+0.1044random-0.5616
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N5421M=-5.021 (flipped)frac_act 0.99 · mean_act +0.020
attr_last-0.0005attr_sum-0.0191attr_max-0.0078Cplus-0.0006Cminus-0.0000empos_contrast-0.0048mask_delta+6.0211random-3.1506
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N8007M=-5.039 (flipped)frac_act 1.00 · mean_act +0.046
attr_last-0.0024attr_sum-0.0022attr_max+0.0025Cplus-0.0024Cminus+0.0000empos_contrast-0.0026mask_delta+6.0392random+0.8930
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N8086M=-0.405 (flipped)frac_act 0.41 · mean_act -0.000
attr_last+0.0000attr_sum+0.0003attr_max+0.0002Cplus-0.0000Cminus-0.0000empos_contrast-0.0001mask_delta+1.4050random+4.7320
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N9095M=+6.995 (amplified)frac_act 0.88 · mean_act +0.004
attr_last-0.0005attr_sum-0.0521attr_max-0.0443Cplus-0.0005Cminus-0.0000empos_contrast-0.0011mask_delta+5.9951random-0.4292
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N9831M=-5.237 (flipped)frac_act 0.97 · mean_act +0.026
attr_last-0.0007attr_sum+0.0140attr_max+0.0142Cplus-0.0007Cminus-0.0000empos_contrast-0.0048mask_delta+6.2371random-2.5402
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N9908M=+7.196 (amplified)frac_act 0.50 · mean_act +0.001
attr_last-0.0008attr_sum+0.0600attr_max+0.0269Cplus-0.0013Cminus-0.0005empos_contrast+0.0044mask_delta+6.1958random-0.4636
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N11784M=+7.313 (amplified)frac_act 0.00 · mean_act -0.105
attr_last+0.0039attr_sum+0.0389attr_max+0.0116Cplus+0.0000Cminus-0.0039empos_contrast+0.0053mask_delta+6.3133random+0.2748
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N12934M=-4.937 (flipped)frac_act 0.57 · mean_act +0.003
attr_last+0.0010attr_sum-0.0058attr_max-0.0094Cplus+0.0011Cminus+0.0001empos_contrast-0.0059mask_delta+5.9368random-0.6393
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N13926M=-5.209 (flipped)frac_act 0.03 · mean_act -0.009
attr_last-0.0001attr_sum+0.0083attr_max-0.0039Cplus+0.0000Cminus+0.0001empos_contrast-0.0020mask_delta+6.2094random-2.1469
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L2·N13949M=-5.827 (flipped)frac_act 0.39 · mean_act -0.003
attr_last+0.0002attr_sum+0.0125attr_max+0.0093Cplus+0.0001Cminus-0.0001empos_contrast-0.0010mask_delta+6.8273random-0.7340
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L3·N6390M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.110
attr_last-0.0113attr_sum+0.6067attr_max+0.1262Cplus+0.0000Cminus+0.0113empos_contrast-0.0006mask_delta+0.0000random-0.0824
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L4·N4152M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.246
attr_last+0.0821attr_sum+0.0151attr_max+0.0535Cplus+0.0821Cminus+0.0000empos_contrast+0.0016mask_delta+0.0000random-1.9081
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 -0.438
attr_last-0.0144attr_sum+0.3281attr_max-0.0128Cplus+0.0000Cminus+0.0144empos_contrast-0.0041mask_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 -0.459
attr_last+0.0102attr_sum-0.7500attr_max-0.5947Cplus+0.0000Cminus-0.0102empos_contrast+0.0012mask_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 -0.436
attr_last-0.0272attr_sum+0.3916attr_max-0.0800Cplus+0.0000Cminus+0.0272empos_contrast+0.0078mask_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 0.40 · mean_act -0.001
attr_last-0.0001attr_sum+0.0002attr_max-0.0022Cplus+0.0001Cminus+0.0002empos_contrast-0.0007mask_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 0.00 · mean_act -0.051
attr_last+0.0013attr_sum+0.0045attr_max-0.0009Cplus+0.0000Cminus-0.0013empos_contrast-0.0021mask_delta+0.0000random-4.4102
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N2134M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.312
attr_last-0.0739attr_sum-0.1138attr_max-0.0751Cplus+0.0000Cminus+0.0739empos_contrast+0.0108mask_delta+0.0000random-1.9808
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N7414M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.584
attr_last-0.0911attr_sum-0.1054attr_max-0.0905Cplus+0.0000Cminus+0.0911empos_contrast-0.0063mask_delta+0.0000random-1.3042
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 +0.516
attr_last+0.1077attr_sum+0.1127attr_max+0.1088Cplus+0.1077Cminus+0.0000empos_contrast-0.0059mask_delta+0.0000random+1.4205
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 0.80 · mean_act +0.009
attr_last+0.0002attr_sum+0.0052attr_max+0.0023Cplus+0.0002Cminus-0.0000empos_contrast+0.0006mask_delta+0.0000random-4.5940
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L12·N1062M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.598
attr_last-0.1397attr_sum-0.1842attr_max-0.1360Cplus-0.1397Cminus+0.0000empos_contrast-0.0744mask_delta+0.0000random+0.2722
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L13·N6149M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.314
attr_last+0.0933attr_sum+0.0953attr_max+0.0933Cplus+0.0933Cminus+0.0000empos_contrast+0.0202mask_delta+0.0000random-1.0560
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 +0.986
attr_last-0.1517attr_sum-0.1976attr_max-0.1519Cplus-0.1517Cminus+0.0000empos_contrast-0.1666mask_delta+0.0000random-0.1297
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L16·N1241M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.558
attr_last-0.0829attr_sum-0.3095attr_max-0.0820Cplus-0.0829Cminus+0.0000empos_contrast-0.0158mask_delta+0.0000random+0.4352
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 0.40 · mean_act -0.003
attr_last-0.0000attr_sum+0.0007attr_max+0.0005Cplus-0.0001Cminus-0.0001empos_contrast+0.0019mask_delta+0.0000random+4.3238
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L24·N8167M=+1.000 (unchanged)frac_act 0.00 · mean_act -1.405
attr_last-0.0622attr_sum-0.0728attr_max-0.0620Cplus+0.0000Cminus+0.0622empos_contrast+0.1064mask_delta+0.0000random+1.5580
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 -0.798
attr_last-0.0862attr_sum-0.1495attr_max-0.0870Cplus+0.0000Cminus+0.0862empos_contrast-0.0923mask_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 0.71 · mean_act +0.015
attr_last-0.0001attr_sum-0.0008attr_max-0.0005Cplus-0.0002Cminus-0.0001empos_contrast-0.0130mask_delta+0.0000random-4.3797
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L26·N3059M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.554
attr_last+0.0799attr_sum+0.1105attr_max+0.0799Cplus+0.0799Cminus+0.0000empos_contrast+0.0324mask_delta+0.0000random+0.9581
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 0.86 · mean_act +0.387
attr_last+0.1604attr_sum+0.1860attr_max+0.1565Cplus+0.1504Cminus-0.0100empos_contrast+0.2864mask_delta+0.0000random+0.4949
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 0.98 · mean_act +0.960
attr_last+0.0931attr_sum+0.1123attr_max+0.0929Cplus+0.0929Cminus-0.0002empos_contrast+0.2576mask_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 0.68 · mean_act +0.006
attr_last-0.0002attr_sum+0.0003attr_max+0.0001Cplus-0.0002Cminus-0.0000empos_contrast-0.0014mask_delta+0.0000random-4.3009
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L28·N11731M=+1.000 (unchanged)frac_act 0.00 · mean_act -2.211
attr_last+0.0903attr_sum+0.0986attr_max+0.0907Cplus+0.0000Cminus-0.0903empos_contrast-0.6869mask_delta+0.0000random-0.1074
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 1.00 · mean_act +2.204
attr_last+0.1232attr_sum+0.1657attr_max+0.1234Cplus+0.1232Cminus+0.0000empos_contrast+0.6022mask_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 -1.976
attr_last+0.2944attr_sum+0.2961attr_max+0.2944Cplus+0.0000Cminus-0.2944empos_contrast-0.0090mask_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 +6.934
attr_last+0.1697attr_sum+0.1367attr_max+0.1697Cplus+0.1697Cminus+0.0000empos_contrast+0.0530mask_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 -1.793
attr_last-0.0654attr_sum-0.0656attr_max-0.0654Cplus+0.0000Cminus+0.0654empos_contrast-0.0321mask_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 0.00 · mean_act -3.784
attr_last-0.4555attr_sum-0.4565attr_max-0.4555Cplus+0.0000Cminus+0.4555empos_contrast-1.1595mask_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 0.50 · mean_act -0.000
attr_last+0.0001attr_sum+0.0001attr_max+0.0001Cplus-0.0000Cminus-0.0001empos_contrast-0.0018mask_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 +2.795
attr_last-0.4967attr_sum-0.4970attr_max-0.4967Cplus-0.4967Cminus+0.0000empos_contrast-0.2175mask_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 1.00 · mean_act +0.721
attr_last+0.1141attr_sum+0.1127attr_max+0.1141Cplus+0.1141Cminus+0.0000empos_contrast+0.1781mask_delta+0.0000random-0.1736
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 0.06 · mean_act -4.368
attr_last+0.0178attr_sum+0.0178attr_max+0.0178Cplus-0.0005Cminus-0.0183empos_contrast+0.5922mask_delta+0.0000random-1.4248
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N5811M=+1.000 (unchanged)frac_act 0.01 · mean_act -0.304
attr_last+0.1364attr_sum+0.1364attr_max+0.1364Cplus+0.0001Cminus-0.1362empos_contrast-0.0329mask_delta+0.0000random-0.3526
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N5835M=+1.000 (unchanged)frac_act 0.35 · mean_act -0.204
attr_last-0.0779attr_sum-0.0779attr_max-0.0779Cplus-0.0081Cminus+0.0697empos_contrast-0.1608mask_delta+0.0000random-0.9731
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 1.00 · mean_act +6.181
attr_last-0.0078attr_sum-0.0078attr_max-0.0078Cplus-0.0078Cminus+0.0000empos_contrast+0.9205mask_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 0.03 · mean_act -0.805
attr_last-0.1409attr_sum-0.1409attr_max-0.1409Cplus+0.0001Cminus+0.1410empos_contrast-0.1400mask_delta+0.0000random+0.3712
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N9187M=+1.000 (unchanged)frac_act 0.50 · mean_act +0.124
attr_last-0.0084attr_sum-0.0084attr_max-0.0084Cplus-0.0081Cminus+0.0003empos_contrast+0.6318mask_delta+0.0000random-1.6372
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N9915M=+1.000 (unchanged)frac_act 0.00 · mean_act -2.993
attr_last+0.0179attr_sum+0.0179attr_max+0.0179Cplus+0.0000Cminus-0.0179empos_contrast-0.5754mask_delta+0.0000random+0.5283
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 0.03 · mean_act -2.125
attr_last-0.0054attr_sum-0.0054attr_max-0.0054Cplus+0.0001Cminus+0.0055empos_contrast+0.8095mask_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 0.41 · mean_act -0.009
attr_last+0.0001attr_sum+0.0001attr_max+0.0001Cplus-0.0001Cminus-0.0003empos_contrast+0.0075mask_delta+0.0000random-4.5677
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N12222M=+1.000 (unchanged)frac_act 0.21 · mean_act -0.485
attr_last+0.0005attr_sum+0.0005attr_max+0.0005Cplus-0.0015Cminus-0.0020empos_contrast+0.5629mask_delta+0.0000random-1.1885
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 +1.775
attr_last+0.1331attr_sum+0.1331attr_max+0.1331Cplus+0.1331Cminus+0.0000empos_contrast+0.1579mask_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 -0.572
attr_last-0.0963attr_sum-0.0963attr_max-0.0963Cplus+0.0000Cminus+0.0963empos_contrast-0.0007mask_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 -3.061
attr_last-1.1021attr_sum-1.1021attr_max-1.1021Cplus+0.0000Cminus+1.1021empos_contrast-0.4940mask_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 0.01 · mean_act -8.393
attr_last+0.0035attr_sum+0.0035attr_max+0.0035Cplus-0.0001Cminus-0.0037empos_contrast+1.1256mask_delta+0.0000random+0.4331
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus