INDEX

Explanations

neutral terms and concepts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 millenn

-0.90

teenth

-0.78

Mill

-0.72

cano

-0.70

Hop

-0.69

omething

-0.69

soDeliveryDate

-0.68

hner

-0.68

 therap

-0.67

FAQ

-0.67

POSITIVE LOGITS

izing

1.27

ization

1.15

izers

1.12

izer

1.11

izes

1.09

ize

1.07

ized

1.07

ity

0.95

izable

0.94

ising

0.90

Activations Density 0.016%

neutral terms and concepts

No Comments

No Known Activations