INDEX

Explanations

prepositions used in phrases related to solutions or answers

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate questions or inquiries about answers

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

hesion

-0.77

cephal

-0.73

chy

-0.70

ONSORED

-0.70

ivities

-0.69

lim

-0.69

outhern

-0.68

kinson

-0.68

uous

-0.66

weeney

-0.65

POSITIVE LOGITS

answer

0.83

 unanswered

0.74

 answer

0.73

 answ

0.73

 trivia

0.72

 constitu

0.69

 answers

0.68

 Answers

0.67

 Tolkien

0.67

 question

0.66

Activations Density 0.117%

prepositions used in phrases related to solutions or answers

phrases that indicate questions or inquiries about answers

No Comments

No Known Activations