INDEX

Explanations

keywords related to providing more information or details about a topic

oai_token-act-pair · gpt-3.5-turbo

references to additional information or details

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

ĸļ

-0.79

ãĥ¯

-0.79

ãĥĵ

-0.74

RN

-0.73

Enlarge

-0.72

mobi

-0.72

ãĥ¼ãĥĨ

-0.71

ãĥ³ãĤ¸

-0.70

ãĥ¼ãĥ«

-0.66

BIT

-0.66

POSITIVE LOGITS

 purposes

0.91

 inquiries

0.81

 information

0.80

 sake

0.78

 info

0.78

 particulars

0.75

 adventurous

0.75

 reasons

0.73

 INFORMATION

0.73

 enqu

0.72

Activations Density 0.059%

keywords related to providing more information or details about a topic

references to additional information or details

No Comments

No Known Activations