INDEX

Explanations

phrases with "as" followed by a noun or pronoun

oai_token-act-pair · gpt-3.5-turbo

phrases that establish roles or functions of entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ubb

-0.72

Rub

-0.67

raq

-0.63

ros

-0.60

"}],"

-0.60

ople

-0.60

zza

-0.59

aceae

-0.59

 COUR

-0.57

jee

-0.57

POSITIVE LOGITS

 opposed

1.11

 well

1.01

pires

0.93

pired

0.92

part

0.90

 follows

0.89

 evidenced

0.86

piring

0.84

 part

0.84

well

0.83

Activations Density 0.138%

phrases with "as" followed by a noun or pronoun

phrases that establish roles or functions of entities

No Comments

No Known Activations