INDEX

Explanations

negative outcomes or situations

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

0:0.01

1:0.01

2:0.08

3:0.06

4:0.11

5:0.02

6:0.03

7:0.41

8:0.04

9:0.03

10:0.08

11:0.05

Negative Logits

hid

-1.80

xus

-1.74

icons

-1.73

nih

-1.65

HIP

-1.60

thumbnails

-1.54

mens

-1.54

ipedia

-1.54

opathy

-1.51

 shielded

-1.51

POSITIVE LOGITS

 sour

1.81

 disappointment

1.64

 bitter

1.56

 Bought

1.50

 disillusion

1.49

terday

1.48

 persuasion

1.46

 mood

1.46

 depleted

1.45

 unlucky

1.45

Activations Density 0.000%

negative outcomes or situations

No Known Activations