INDEX

Explanations

mentions of the word "rogue"

oai_token-act-pair · gpt-3.5-turbo

instances of the substring "rog"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Â¢

-0.72

 Speedway

-0.71

AAP

-0.69

 Claus

-0.69

Fas

-0.68

éĹĺ

-0.65

 USPS

-0.61

++++

-0.60

 Faul

-0.59

 birth

-0.58

POSITIVE LOGITS

rog

1.48

raphic

1.24

aming

1.09

raphics

1.08

eln

0.98

raph

0.98

allery

0.92

rams

0.92

uild

0.92

otypes

0.91

Activations Density 0.003%

mentions of the word "rogue"

instances of the substring "rog"

No Comments

No Known Activations