INDEX

Explanations

words related to negative emotions, particularly anger

oai_token-act-pair · gpt-3.5-turbo

expressions of anger and related emotions

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

•layer 5 - anger

No Comments

Negative Logits

artifacts

-0.76

glas

-0.74

 livest

-0.72

icrobial

-0.71

ramer

-0.69

uchin

-0.68

 Vide

-0.68

akeru

-0.66

MER

-0.66

nown

-0.64

POSITIVE LOGITS

 anger

0.85

rained

0.83

 indignation

0.82

fulness

0.82

wart

0.81

ful

0.80

ingly

0.79

FUL

0.79

 resentment

0.76

bird

0.75

Activations Density 0.020%

words related to negative emotions, particularly anger

expressions of anger and related emotions

No Comments

No Known Activations