INDEX

Explanations

instances of phrases indicating political correctness or inclusivity

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

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

Not in Any Lists

No Comments

Head Attr Weights

0:0.08

1:0.11

2:0.03

3:0.04

4:0.03

5:0.39

6:0.04

7:0.02

8:0.05

9:0.07

10:0.05

11:0.03

Negative Logits

assetsadobe

-2.13

 Marketable

-1.89

 Manit

-1.75

 istg

-1.71

..........

-1.70

SAN

-1.69

 locker

-1.67

[&

-1.65

 sqor

-1.62

SK

-1.60

POSITIVE LOGITS

�

2.08

cone

1.92

anything

1.87

asy

1.83

argument

1.81

�

1.78

authorized

1.77

too

1.74

�

1.72

echo

1.71

Activations Density 0.123%

instances of phrases indicating political correctness or inclusivity

No Comments

No Known Activations