INDEX

Explanations

references to the country of India

oai_token-act-pair · gpt-3.5-turbo

mentions of "Ind" or indicators associated with "Ind."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 compr

-0.95

EStream

-0.89

sonian

-0.81

thodox

-0.78

avorite

-0.73

TION

-0.71

xit

-0.68

abwe

-0.68

RTX

-0.66

 litter

-0.66

POSITIVE LOGITS

irect

1.19

ividually

1.16

ebted

1.07

enture

1.04

icative

1.03

icator

1.01

epend

0.98

icates

0.94

ivid

0.92

ications

0.91

Activations Density 0.009%

references to the country of India

mentions of "Ind" or indicators associated with "Ind."

No Comments

No Known Activations