INDEX

Explanations

keywords related to writing or reporting

oai_token-act-pair · gpt-3.5-turbo

references to various topics and subjects being discussed or covered in writing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 teasp

-0.91

©¶æ¥µ

-0.80

ngth

-0.76

chers

-0.74

vertisement

-0.73

osponsors

-0.72

ridor

-0.72

Pwr

-0.71

atten

-0.70

ilant

-0.69

POSITIVE LOGITS

 topics

1.75

 subjects

1.30

 matters

1.20

 issues

1.19

 everything

1.16

 themes

1.16

 politics

1.14

 sexuality

1.13

 topical

1.11

 economics

1.11

Activations Density 0.222%

keywords related to writing or reporting

references to various topics and subjects being discussed or covered in writing

No Comments

No Known Activations