INDEX

Explanations

references to specific individuals and their actions or attributes, particularly in contexts related to life situations, comparisons, or evaluations

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

Head Attr Weights

0:0.01

1:0.01

2:0.15

3:0.27

4:0.19

5:0.03

6:0.03

7:0.07

8:0.04

9:0.03

10:0.06

11:0.06

Negative Logits

��極

-2.18

APD

-1.79

%);

-1.62

%).

-1.57

 awaits

-1.48

 Calls

-1.45

ighting

-1.44

WARNING

-1.43

interstitial

-1.42

Adds

-1.41

POSITIVE LOGITS

?",

1.57

 tomorrow

1.49

 then

1.46

 somebody

1.44

1.41

 someday

1.41

 yesterday

1.41

 hypot

1.40

 THEN

1.40

?"

1.39

Activations Density 0.105%

references to specific individuals and their actions or attributes, particularly in contexts related to life situations, comparisons, or evaluations

No Known Activations

references to specific individuals and their actions or attributes, particularly in contexts related to life situations, comparisons, or evaluations

No Known Activations