INDEX

Explanations

mentions or abbreviations related to news sources or organizations

oai_token-act-pair · gpt-3.5-turbo

references to national security

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

fare

-0.73

Äĩ

-0.72

======

-0.65

zona

-0.62

thumbnails

-0.62

verages

-0.61

 mats

-0.60

ments

-0.60

 otherwise

-0.60

Ã§a

-0.59

POSITIVE LOGITS

FW

1.26

fw

1.12

ertodd

1.01

ettings

0.85

erve

0.82

lÃ©

0.82

erver

0.81

daq

0.81

ensitive

0.81

ession

0.80

Activations Density 0.018%

mentions or abbreviations related to news sources or organizations

references to national security

No Comments

No Known Activations