INDEX

Explanations

words related to opposition or conflict

oai_token-act-pair · gpt-3.5-turbo

references to opposing viewpoints or forces

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

onics

-0.74

seed

-0.70

ju

-0.69

Kev

-0.69

oner

-0.69

oled

-0.68

ikan

-0.67

ager

-0.66

del

-0.66

atche

-0.66

POSITIVE LOGITS

 opposing

1.30

 opponents

0.96

 oppos

0.93

 viewpoints

0.90

 opponent

0.89

 undermin

0.85

 opposition

0.83

 foes

0.82

 sides

0.81

 strugg

0.81

Activations Density 0.014%

words related to opposition or conflict

references to opposing viewpoints or forces

No Comments

No Known Activations