INDEX

Explanations

phrases related to legal and bureaucratic processes

oai_token-act-pair · gpt-3.5-turbo

references to legal or procedural processes

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

 Flavoring

-0.71

vation

-0.64

 Hots

-0.63

 Haram

-0.60

hiro

-0.60

oft

-0.59

 earnest

-0.57

mson

-0.57

 Portug

-0.57

 Wand

-0.56

POSITIVE LOGITS

ional

1.14

ions

0.79

ioned

0.79

 overseen

0.74

DragonMagazine

0.74

 process

0.73

hene

0.71

anus

0.69

wright

0.68

matically

0.66

Activations Density 0.032%

phrases related to legal and bureaucratic processes

references to legal or procedural processes

No Comments

No Known Activations