INDEX

Explanations

organizations or entities and the roles or positions associated with them

oai_token-act-pair · gpt-3.5-turbo

the word "which" used in a context of explanation or clarification

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

Behind

-0.74

athi

-0.69

grim

-0.66

ve

-0.63

Ott

-0.63

WARE

-0.62

Adams

-0.61

Gy

-0.61

ben

-0.61

Bas

-0.60

POSITIVE LOGITS

 operates

0.88

 comprises

0.87

 specializes

0.87

 translates

0.86

 oversees

0.84

 encompasses

0.82

 resulted

0.82

 governs

0.81

 consists

0.81

 originated

0.80

Activations Density 0.132%

organizations or entities and the roles or positions associated with them

the word "which" used in a context of explanation or clarification

No Comments

No Known Activations