INDEX

Explanations

words related to opposition or conflict

oai_token-act-pair · gpt-3.5-turbo

instances of the word "opposing" in various contexts

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

atche

-0.78

seed

-0.75

nesia

-0.70

oled

-0.70

Redditor

-0.69

onics

-0.67

mberg

-0.67

agna

-0.67

ebook

-0.66

ersen

-0.65

POSITIVE LOGITS

 opposing

1.10

 opponents

0.93

 viewpoints

0.92

 oppos

0.88

 undermin

0.88

 opponent

0.88

 sides

0.86

 foes

0.83

 enemy

0.80

 combatants

0.77

Activations Density 0.006%

words related to opposition or conflict

instances of the word "opposing" in various contexts

No Comments

No Known Activations