INDEX

Explanations

references to a specific person's name, specifically "Lou."

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

Constructed

-0.83

flies

-0.73

ãģĨ

-0.67

 Tactics

-0.66

CoC

-0.65

 fences

-0.63

é¾į

-0.62

 fence

-0.62

conservancy

-0.61

AQ

-0.60

POSITIVE LOGITS

vre

1.21

Geh

0.98

ving

0.90

ochet

0.86

illard

0.86

rie

0.86

zzy

0.85

der

0.83

lette

0.82

utenant

0.80

Activations Density 0.021%

references to a specific person's name, specifically "Lou."

No Comments

No Known Activations