INDEX

Explanations

references to casualties and operational losses related to military conflicts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

DragonMagazine

-0.93

hat

-0.74

learning

-0.74

Rum

-0.67

onomy

-0.66

rift

-0.64

Hat

-0.64

hood

-0.63

learn

-0.63

 misunderstanding

-0.63

POSITIVE LOGITS

 consecut

0.76

 outright

0.75

 apiece

0.73

 classified

0.73

pointers

0.70

 clustered

0.69

 categorized

0.69

 individually

0.68

 handguns

0.67

 deemed

0.67

Activations Density 0.132%

references to casualties and operational losses related to military conflicts

No Comments

No Known Activations