INDEX

Explanations

words related to competitive interactions or conflict, especially in a gaming context

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

olia

-0.81

vest

-0.73

 Sabha

-0.71

acca

-0.70

olyn

-0.69

ocratic

-0.67

oots

-0.66

gres

-0.65

mberg

-0.65

Balt

-0.64

POSITIVE LOGITS

onent

0.90

 batters

0.90

 opponent

0.83

 combatants

0.74

foe

0.73

 opponents

0.70

hesis

0.69

 hitters

0.68

 batter

0.67

 strugg

0.65

Activations Density 0.035%

words related to competitive interactions or conflict, especially in a gaming context

No Comments

No Known Activations