InterpUpdates› MCQ attribution + mask inspection

MCQ single-token attribution + mask inspection

Config mask_L6_lam3e-5 · EM 0.09% (terra) · checkpoint mask_L6_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 336 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 (L6)also in untrained base top-20
attr_lastL5, L13, L15, L16, L18, L26, L30, L310/106/10
attr_sumL3, L5, L9, L11, L12, L13, L16, L18…0/107/10
attr_maxL0, L1, L3, L4, L9, L12, L16, L30…0/106/10
top_by_CplusL12, L13, L16, L26, L27, L29, L30, L310/107/10
top_by_CminusL5, L7, L13, L14, L20, L24, L30, L310/108/10
empos_contrastL310/101/10
mask_deltaL610/100/10
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·N13006-0.8781—
2L30·N8560-0.2566—
3L30·N4803-0.2499—
4L16·N10990+0.1867—
5L5·N7012-0.1823—
6L26·N3526+0.1735—
7L15·N8286-0.1699—
8L13·N4208-0.1665—
9L16·N9215-0.1635—
10L18·N7417-0.1603—

attr_sum top 10 · grad×act summed over positions

#neuronattr_sumMauto-interp description
1L31·N13006-0.8781—
2L3·N6390+0.6343—
3L9·N4255-0.5070—
4L5·N7012+0.4845—
5L13·N4208-0.3126—
6L12·N13860-0.3102—
7L11·N11321-0.3020—
8L18·N7417-0.2911—
9L16·N1241-0.2724—
10L30·N8560-0.2555—

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

#neuronattr_maxMauto-interp description
1L31·N13006-0.8781—
2L30·N8560-0.2566—
3L30·N4803-0.2499—
4L12·N10099+0.2422—
5L0·N12387-0.2263—
6L4·N8564+0.2092—
7L3·N11598-0.1981—
8L1·N7179+0.1946—
9L16·N10990+0.1903—
10L9·N4255-0.1855—

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

#neuronCplusMauto-interp description
1L16·N10990+0.1867—
2L30·N3382+0.1564—
3L29·N1141+0.1247—
4L26·N3526+0.1220—
5L31·N12721+0.1219—
6L16·N10564+0.0906—
7L30·N13562+0.0835—
8L12·N8856+0.0817—
9L27·N6317+0.0807—
10L13·N6149+0.0731—

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

#neuronCminusMauto-interp description
1L31·N13006+0.8781—
2L30·N4803+0.2491—
3L5·N7012+0.1823—
4L13·N4208+0.1665—
5L7·N6673+0.0970—
6L31·N8120+0.0836—
7L31·N5835+0.0821—
8L20·N4299+0.0798—
9L14·N13181+0.0680—
10L24·N8167+0.0653—

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

#neuronempos_contrastMauto-interp description
1L31·N1209+1.7641—
2L31·N311+1.2322—
3L31·N13336+1.2022—
4L31·N2564-1.1796—
5L31·N2398+0.8305—
6L31·N5919-0.6486—
7L31·N6614+0.6227—
8L31·N12540+0.6209—
9L31·N5540+0.6135—
10L31·N2344+0.6007—

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

#neuronmask_deltaMauto-interp description
1L6·N2110+7.5382—
2L6·N655+7.5024—
3L6·N4747+7.0133—
4L6·N354+6.9541—
5L6·N7943+6.6909—
6L6·N9808+6.6536—
7L6·N7757+6.6483—
8L6·N3560+6.5946—
9L6·N5595+6.5832—
10L6·N5618+6.4721—

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 64 unique neurons, each rendered once

L0·N12387M=+1.000 (unchanged)frac_act 0.26 · mean_act -0.000
attr_last-0.0000attr_sum-0.2259attr_max-0.2263Cplus+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·N7179M=+1.000 (unchanged)frac_act 0.06 · mean_act -0.001
attr_last+0.0000attr_sum+0.2098attr_max+0.1946Cplus+0.0000Cminus-0.0000empos_contrast+0.0002mask_delta+0.0000random+0.0362
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 0.24 · mean_act -0.002
attr_last+0.0005attr_sum-0.0001attr_max+0.0006Cplus-0.0000Cminus-0.0005empos_contrast-0.0007mask_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 0.56 · mean_act +0.000
attr_last+0.0001attr_sum-0.0008attr_max-0.0010Cplus+0.0001Cminus-0.0000empos_contrast+0.0000mask_delta+0.0000random+4.7320
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 -0.034
attr_last+0.0012attr_sum+0.6343attr_max+0.0319Cplus+0.0000Cminus-0.0012empos_contrast-0.0025mask_delta+0.0000random-0.0824
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L3·N11598M=+1.000 (unchanged)frac_act 0.10 · mean_act -0.004
attr_last+0.0002attr_sum-0.2151attr_max-0.1981Cplus+0.0000Cminus-0.0002empos_contrast-0.0009mask_delta+0.0000random+0.7609
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L4·N8564M=+1.000 (unchanged)frac_act 0.90 · mean_act +0.020
attr_last-0.0000attr_sum+0.2286attr_max+0.2092Cplus-0.0001Cminus-0.0000empos_contrast+0.0055mask_delta+0.0000random+1.3991
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 -0.409
attr_last-0.1823attr_sum+0.4845attr_max+0.1725Cplus+0.0000Cminus+0.1823empos_contrast-0.0023mask_delta+0.0000random+1.8313
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N354M=-5.954 (flipped)frac_act 0.59 · mean_act +0.010
attr_last-0.0001attr_sum-0.0227attr_max-0.0122Cplus-0.0003Cminus-0.0003empos_contrast+0.0036mask_delta+6.9541random-1.1039
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N655M=+8.502 (amplified)frac_act 0.15 · mean_act -0.047
attr_last+0.0005attr_sum+0.0475attr_max+0.0206Cplus+0.0000Cminus-0.0004empos_contrast-0.0244mask_delta+7.5024random+0.1780
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N2110M=-6.538 (flipped)frac_act 0.00 · mean_act -0.287
attr_last-0.0102attr_sum-0.0425attr_max-0.0155Cplus+0.0000Cminus+0.0102empos_contrast-0.0288mask_delta+7.5382random+1.2952
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N3560M=-5.595 (flipped)frac_act 0.08 · mean_act -0.071
attr_last+0.0011attr_sum+0.0344attr_max+0.0029Cplus-0.0000Cminus-0.0011empos_contrast-0.0109mask_delta+6.5946random-1.2670
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N4747M=-6.013 (flipped)frac_act 1.00 · mean_act +0.265
attr_last+0.0108attr_sum+0.0719attr_max+0.0143Cplus+0.0108Cminus+0.0000empos_contrast+0.0102mask_delta+7.0133random+2.0900
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N5595M=-5.583 (flipped)frac_act 0.00 · mean_act -0.128
attr_last+0.0070attr_sum-0.0403attr_max-0.0046Cplus+0.0000Cminus-0.0070empos_contrast+0.0039mask_delta+6.5832random-2.0372
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N5618M=+7.472 (amplified)frac_act 0.40 · mean_act -0.021
attr_last-0.0006attr_sum-0.0056attr_max-0.0006Cplus-0.0001Cminus+0.0005empos_contrast+0.0545mask_delta+6.4721random-0.0717
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N7757M=+7.648 (amplified)frac_act 1.00 · mean_act +0.156
attr_last-0.0081attr_sum-0.0030attr_max+0.0034Cplus-0.0081Cminus+0.0000empos_contrast-0.0007mask_delta+6.6483random+0.0993
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N7943M=+7.691 (amplified)frac_act 0.93 · mean_act +0.057
attr_last-0.0004attr_sum-0.0328attr_max-0.0216Cplus-0.0004Cminus+0.0000empos_contrast+0.0046mask_delta+6.6909random-1.8911
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L6·N9808M=-5.654 (flipped)frac_act 0.56 · mean_act +0.001
attr_last+0.0001attr_sum+0.0301attr_max+0.0478Cplus+0.0004Cminus+0.0003empos_contrast-0.0019mask_delta+6.6536random+0.2012
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 -0.420
attr_last-0.0970attr_sum+0.1933attr_max-0.1725Cplus+0.0000Cminus+0.0970empos_contrast-0.0013mask_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 0.90 · mean_act +0.006
attr_last-0.0000attr_sum+0.0031attr_max+0.0018Cplus-0.0000Cminus+0.0000empos_contrast+0.0001mask_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 0.08 · mean_act -0.013
attr_last-0.0001attr_sum+0.0036attr_max-0.0023Cplus-0.0000Cminus+0.0001empos_contrast-0.0006mask_delta+0.0000random-4.4102
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L9·N4255M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.435
attr_last-0.0991attr_sum-0.5070attr_max-0.1855Cplus-0.0991Cminus+0.0000empos_contrast+0.0023mask_delta+0.0000random+0.7887
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 +0.567
attr_last+0.0130attr_sum-0.3020attr_max-0.1097Cplus+0.0130Cminus+0.0000empos_contrast-0.0032mask_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 0.92 · mean_act +0.014
attr_last-0.0018attr_sum+0.0067attr_max+0.0008Cplus-0.0018Cminus-0.0000empos_contrast-0.0015mask_delta+0.0000random-4.5940
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L12·N8856M=+1.000 (unchanged)frac_act 0.82 · mean_act +0.089
attr_last+0.0815attr_sum+0.1008attr_max+0.0816Cplus+0.0817Cminus+0.0002empos_contrast+0.0094mask_delta+0.0000random+1.0835
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 +0.310
attr_last-0.0074attr_sum+0.1536attr_max+0.2422Cplus-0.0074Cminus+0.0000empos_contrast-0.0156mask_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 +0.616
attr_last-0.1193attr_sum-0.3102attr_max-0.1136Cplus-0.1193Cminus+0.0000empos_contrast-0.0117mask_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 -0.772
attr_last-0.1665attr_sum-0.3126attr_max-0.1672Cplus+0.0000Cminus+0.1665empos_contrast+0.0072mask_delta+0.0000random-0.4598
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 +0.137
attr_last+0.0731attr_sum+0.0685attr_max+0.0733Cplus+0.0731Cminus+0.0000empos_contrast+0.0095mask_delta+0.0000random-1.0560
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L14·N13181M=+1.000 (unchanged)frac_act 0.00 · mean_act -0.741
attr_last-0.0680attr_sum-0.0775attr_max-0.0675Cplus+0.0000Cminus+0.0680empos_contrast+0.0347mask_delta+0.0000random+0.4228
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 +0.974
attr_last-0.1699attr_sum-0.1885attr_max-0.1698Cplus-0.1699Cminus+0.0000empos_contrast-0.0983mask_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 +0.528
attr_last-0.0835attr_sum-0.2724attr_max-0.0574Cplus-0.0835Cminus+0.0000empos_contrast-0.0177mask_delta+0.0000random+0.4352
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L16·N9215M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.453
attr_last-0.1635attr_sum-0.1713attr_max-0.1636Cplus-0.1635Cminus+0.0000empos_contrast+0.0366mask_delta+0.0000random+0.1499
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L16·N10564M=+1.000 (unchanged)frac_act 0.73 · mean_act +0.083
attr_last+0.0930attr_sum+0.0931attr_max+0.0917Cplus+0.0906Cminus-0.0024empos_contrast+0.0434mask_delta+0.0000random+0.6021
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L16·N10990M=+1.000 (unchanged)frac_act 1.00 · mean_act +2.668
attr_last+0.1867attr_sum+0.2160attr_max+0.1903Cplus+0.1867Cminus+0.0000empos_contrast+0.1262mask_delta+0.0000random-0.4357
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L18·N7417M=+1.000 (unchanged)frac_act 1.00 · mean_act +0.606
attr_last-0.1603attr_sum-0.2911attr_max-0.1695Cplus-0.1603Cminus+0.0000empos_contrast+0.0083mask_delta+0.0000random+0.9041
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 0.61 · mean_act +0.002
attr_last-0.0000attr_sum+0.0006attr_max+0.0002Cplus-0.0000Cminus+0.0000empos_contrast+0.0021mask_delta+0.0000random+4.3238
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L20·N4299M=+1.000 (unchanged)frac_act 0.01 · mean_act -0.316
attr_last-0.0793attr_sum-0.1065attr_max-0.0811Cplus+0.0005Cminus+0.0798empos_contrast+0.0291mask_delta+0.0000random-0.0814
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 -1.396
attr_last-0.0653attr_sum-0.0717attr_max-0.0652Cplus+0.0000Cminus+0.0653empos_contrast+0.0614mask_delta+0.0000random+1.5580
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 0.41 · mean_act -0.007
attr_last+0.0001attr_sum-0.0002attr_max-0.0001Cplus-0.0001Cminus-0.0002empos_contrast-0.0024mask_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 0.62 · mean_act +0.164
attr_last+0.1735attr_sum+0.1822attr_max+0.1695Cplus+0.1220Cminus-0.0515empos_contrast+0.1778mask_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 +0.856
attr_last+0.0810attr_sum+0.0900attr_max+0.0807Cplus+0.0807Cminus-0.0003empos_contrast+0.0859mask_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 0.41 · mean_act -0.003
attr_last+0.0001attr_sum+0.0000attr_max-0.0001Cplus-0.0001Cminus-0.0002empos_contrast-0.0016mask_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 0.98 · mean_act +1.724
attr_last+0.1256attr_sum+0.1592attr_max+0.1238Cplus+0.1247Cminus-0.0010empos_contrast+0.1622mask_delta+0.0000random-0.5724
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 +6.539
attr_last+0.1564attr_sum+0.1284attr_max+0.1563Cplus+0.1564Cminus+0.0000empos_contrast-0.1429mask_delta+0.0000random+0.9214
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 -2.465
attr_last-0.2499attr_sum-0.2502attr_max-0.2499Cplus-0.0008Cminus+0.2491empos_contrast-0.2432mask_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 0.54 · mean_act -0.004
attr_last+0.0004attr_sum+0.0004attr_max+0.0004Cplus+0.0001Cminus-0.0003empos_contrast-0.0028mask_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 +2.100
attr_last-0.2566attr_sum-0.2555attr_max-0.2566Cplus-0.2566Cminus+0.0000empos_contrast-0.0296mask_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 0.99 · mean_act +0.571
attr_last+0.0843attr_sum+0.0831attr_max+0.0843Cplus+0.0835Cminus-0.0007empos_contrast+0.0970mask_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 0.40 · mean_act -0.338
attr_last+0.0025attr_sum+0.0025attr_max+0.0025Cplus-0.0025Cminus-0.0049empos_contrast+1.2322mask_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 0.33 · mean_act -1.130
attr_last+0.0071attr_sum+0.0071attr_max+0.0071Cplus-0.0018Cminus-0.0089empos_contrast+1.7641mask_delta+0.0000random-1.4248
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N2344M=+1.000 (unchanged)frac_act 0.04 · mean_act -1.143
attr_last-0.0087attr_sum-0.0087attr_max-0.0087Cplus+0.0001Cminus+0.0088empos_contrast+0.6007mask_delta+0.0000random-1.1101
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N2398M=+1.000 (unchanged)frac_act 1.00 · mean_act +8.505
attr_last-0.0220attr_sum-0.0220attr_max-0.0220Cplus-0.0220Cminus+0.0000empos_contrast+0.8305mask_delta+0.0000random+0.3985
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 0.09 · mean_act -4.175
attr_last+0.0052attr_sum+0.0052attr_max+0.0052Cplus-0.0002Cminus-0.0055empos_contrast-1.1796mask_delta+0.0000random-1.2506
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N5540M=+1.000 (unchanged)frac_act 0.80 · mean_act +0.641
attr_last+0.0035attr_sum+0.0035attr_max+0.0035Cplus+0.0033Cminus-0.0002empos_contrast+0.6135mask_delta+0.0000random+0.5838
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 0.07 · mean_act -0.407
attr_last-0.0850attr_sum-0.0850attr_max-0.0850Cplus-0.0029Cminus+0.0821empos_contrast-0.0710mask_delta+0.0000random-0.9731
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N5919M=+1.000 (unchanged)frac_act 0.64 · mean_act +0.367
attr_last-0.0009attr_sum-0.0009attr_max-0.0009Cplus-0.0010Cminus-0.0001empos_contrast-0.6486mask_delta+0.0000random-1.6519
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N6614M=+1.000 (unchanged)frac_act 1.00 · mean_act +3.696
attr_last-0.0011attr_sum-0.0011attr_max-0.0011Cplus-0.0011Cminus+0.0000empos_contrast+0.6227mask_delta+0.0000random-0.7782
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 0.00 · mean_act -0.795
attr_last-0.0836attr_sum-0.0836attr_max-0.0836Cplus-0.0000Cminus+0.0836empos_contrast+0.0545mask_delta+0.0000random+0.3712
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 0.17 · mean_act -0.037
attr_last-0.0000attr_sum-0.0000attr_max-0.0000Cplus-0.0001Cminus-0.0001empos_contrast+0.0171mask_delta+0.0000random-4.5677
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus

L31·N12540M=+1.000 (unchanged)frac_act 0.64 · mean_act +0.330
attr_last-0.0000attr_sum-0.0000attr_max-0.0000Cplus+0.0002Cminus+0.0003empos_contrast+0.6209mask_delta+0.0000random-0.5283
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 +1.595
attr_last+0.1219attr_sum+0.1219attr_max+0.1219Cplus+0.1219Cminus+0.0000empos_contrast+0.0573mask_delta+0.0000random+1.5212
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 -2.832
attr_last-0.8781attr_sum-0.8781attr_max-0.8781Cplus+0.0000Cminus+0.8781empos_contrast-0.3291mask_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 -6.331
attr_last+0.0008attr_sum+0.0008attr_max+0.0008Cplus-0.0004Cminus-0.0012empos_contrast+1.2022mask_delta+0.0000random+0.4331
General corpus (C4 web text)

no data for this corpus

MCQ probe prompts

no data for this corpus