INDEX

Explanations

the word "a" in contexts associated with political discourse or discussions

New Auto-Interp

Top Features by Cosine Similarity

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

0:0.09

1:0.06

2:0.08

3:0.08

4:0.07

5:0.08

6:0.08

7:0.08

8:0.08

9:0.07

10:0.09

11:0.07

Negative Logits

pel

-1.95

oult

-1.95

quickShipAvailable

-1.95

laughs

-1.92

lier

-1.85

mite

-1.75

?).

-1.73

kat

-1.71

mic

-1.67

?),

-1.67

POSITIVE LOGITS

 Starts

1.69

 Chapters

1.67

 Areas

1.65

 Identification

1.65

 Sect

1.62

 Persons

1.60

 Letters

1.57

 Corridor

1.56

 Advisory

1.55

 Confederation

1.55

Activations Density 0.000%

the word "a" in contexts associated with political discourse or discussions

No Known Activations