INDEX

Explanations

phrases related to opposition or contrary viewpoints

oai_token-act-pair · gpt-3.5-turbo

references to opposition or opponents in a discussion or argument

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

Interstitial

-0.76

ä¹ĭ

-0.71

ãģ®éŃĶ

-0.70

çīĪ

-0.69

enegger

-0.68

ãĤ·ãĥ£

-0.65

è¯

-0.63

negie

-0.63

çĶ°

-0.62

WAYS

-0.62

POSITIVE LOGITS

ortun

1.55

onent

1.48

osite

1.44

osition

1.37

onents

1.32

osed

1.05

osing

1.05

ressive

1.01

ressed

0.97

ression

0.96

Activations Density 0.008%

phrases related to opposition or contrary viewpoints

references to opposition or opponents in a discussion or argument

No Comments

No Known Activations