INDEX

Explanations

language related to negative emotions and conflicts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "in" indicating conflict or tension

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 caut

-0.65

chell

-0.65

aunder

-0.63

HAEL

-0.59

 avail

-0.59

 shalt

-0.58

essor

-0.58

onym

-0.57

USB

-0.57

eas

-0.56

POSITIVE LOGITS

effic

1.32

efficiency

1.28

 regards

1.26

roads

1.23

 favor

1.12

 lieu

1.11

 relation

1.10

 spite

1.06

 regard

1.03

ordinate

1.02

Activations Density 0.408%

language related to negative emotions and conflicts

instances of the word "in" indicating conflict or tension

No Comments

No Known Activations