INDEX

Explanations

words related to the status or condition of being operational or functional

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.03

2:0.33

3:0.06

4:0.05

5:0.03

6:0.08

7:0.11

8:0.04

9:0.04

10:0.08

11:0.08

Negative Logits

turned

-1.82

 Himself

-1.60

vered

-1.53

 alone

-1.51

 ripe

-1.50

skilled

-1.45

 doubled

-1.44

 twice

-1.43

 mature

-1.43

 committed

-1.42

POSITIVE LOGITS

aleb

1.85

 corrid

1.79

abin

1.66

bilt

1.60

ieft

1.59

 Cosponsors

1.53

andel

1.53

icro

1.51

 annotation

1.48

olor

1.47

Activations Density 0.001%

words related to the status or condition of being operational or functional

No Comments

No Known Activations