INDEX

Explanations

technical specifications and details related to products, processes, or data analysis

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

sis

-0.71

aver

-0.68

hatt

-0.68

iev

-0.68

Enlarge

-0.64

turn

-0.63

esa

-0.63

mother

-0.62

nown

-0.62

gamer

-0.62

POSITIVE LOGITS

 plenty

1.07

 considerable

1.07

 tons

1.06

an

1.05

1.04

 extensive

1.02

 ample

0.99

 some

0.98

 lots

0.97

 substantial

0.94

Activations Density 5.937%

technical specifications and details related to products, processes, or data analysis

No Comments

No Known Activations