INDEX

Explanations

technical information such as IDs, codes, and descriptions that have a consistent pattern in their formatting

oai_token-act-pair · gpt-3.5-turbo

indicators of online content or posts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Standing

-0.71

 perk

-0.68

 veter

-0.63

Bye

-0.62

 Pound

-0.62

 utilities

-0.62

 perks

-0.61

 routines

-0.60

 Sunny

-0.60

 rewards

-0.60

POSITIVE LOGITS

Posted

0.89

Subject

0.86

Registered

0.85

isine

0.85

English

0.85

Edited

0.84

Updated

0.81

à¦

0.80

abulary

0.80

att

0.79

Activations Density 0.067%

technical information such as IDs, codes, and descriptions that have a consistent pattern in their formatting

indicators of online content or posts

No Comments

No Known Activations