INDEX

Explanations

names of organizations or political entities

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

|--

-0.60

 Leica

-0.54

MSI

-0.53

 conductor

-0.52

 myster

-0.52

 wrench

-0.52

)=(

-0.51

 groin

-0.51

"+

-0.50

POSITIVE LOGITS

ratom

0.86

opia

0.80

 Syndrome

0.80

ploy

0.79

's

0.73

oppers

0.70

utical

0.69

phia

0.69

Mania

0.68

Wiki

0.67

Activations Density 0.625%

names of organizations or political entities

No Comments

No Known Activations