INDEX

Explanations

references to legislative bills and laws

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

 Lines

-0.70

 Assassins

-0.67

 Demons

-0.66

 Distance

-0.64

tsky

-0.63

 Awards

-0.62

 Disorders

-0.62

 Syndrome

-0.61

 endurance

-0.61

 Continent

-0.61

POSITIVE LOGITS

 authorizing

1.08

 legalizing

1.03

 restricting

0.95

 aimed

0.95

 prohibiting

0.93

 proposing

0.93

 banning

0.92

 requiring

0.91

 allowing

0.90

 regulating

0.90

Activations Density 0.080%

references to legislative bills and laws

No Comments

No Known Activations