INDEX

Explanations

terms related to regulations and laws, including categories like race, gender, religion, and national origin

oai_token-act-pair · gpt-3.5-turbo

items and characteristics that are typically listed or categorized

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

herent

-0.70

¬¼

-0.69

ame

-0.64

ername

-0.61

irlf

-0.61

icol

-0.60

iaries

-0.59

 fres

-0.58

ocese

-0.58

ugi

-0.57

POSITIVE LOGITS

etc

1.04

 albeit

0.95

etc

0.74

 namely

0.69

 which

0.69

 however

0.69

 whereas

0.66

 although

0.66

but

0.65

Activations Density 0.796%

terms related to regulations and laws, including categories like race, gender, religion, and national origin

items and characteristics that are typically listed or categorized

No Comments

No Known Activations