INDEX

Explanations

phrases related to military operations

oai_token-act-pair · gpt-3.5-turbo

references to specific operations or missions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

axter

-0.83

ensing

-0.81

bery

-0.81

erers

-0.78

hus

-0.77

iard

-0.74

warts

-0.72

some

-0.72

yards

-0.69

ering

-0.69

POSITIVE LOGITS

 Operation

0.91

 Twist

0.88

 Arrow

0.84

olon

0.83

Operation

0.80

 Pillar

0.76

 Glad

0.74

 Bunker

0.71

 Restore

0.71

 Mongo

0.71

Activations Density 0.015%

phrases related to military operations

references to specific operations or missions

No Comments

No Known Activations