INDEX

Explanations

phrases related to closeness, accuracy, goodness, and bigness

oai_token-act-pair · gpt-3.5-turbo

adjectives and adverbs indicating quality, proximity, and accuracy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

istries

-0.80

vation

-0.76

othal

-0.72

 Bless

-0.67

 Vacc

-0.65

oscope

-0.65

throw

-0.64

emen

-0.64

 Carnival

-0.64

ardo

-0.63

POSITIVE LOGITS

 edged

0.79

 replacements

0.71

 indent

0.70

 tong

0.70

ifted

0.69

 ascended

0.69

 progresses

0.68

 learners

0.68

 dared

0.68

 divid

0.68

Activations Density 0.114%

phrases related to closeness, accuracy, goodness, and bigness

adjectives and adverbs indicating quality, proximity, and accuracy

No Comments

No Known Activations