INDEX

Explanations

mentions of tanks

oai_token-act-pair · gpt-3.5-turbo

references to tanks or tank-related topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

âĸ¬

-0.87

theless

-0.70

çīĪ

-0.66

 Brotherhood

-0.63

————

-0.63

éĹ

-0.63

BOOK

-0.63

Lod

-0.63

 DMCA

-0.61

 Verd

-0.60

POSITIVE LOGITS

erness

1.19

yard

1.10

mates

1.00

tank

0.99

ards

0.97

ard

0.97

buster

0.94

roller

0.93

ered

0.90

ering

0.89

Activations Density 0.025%

mentions of tanks

references to tanks or tank-related topics

No Comments

No Known Activations