INDEX

Explanations

occurrences of the word "to" with various contexts and phrases

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ãĤ¦ãĤ¹

-0.66

opens

-0.65

noticed

-0.65

bats

-0.65

blooded

-0.64

imore

-0.60

ashes

-0.60

apps

-0.59

cart

-0.59

Apps

-0.59

POSITIVE LOGITS

be

1.05

 embody

1.00

 indicate

0.95

 have

0.94

 endorse

0.93

 originate

0.92

 belong

0.91

 defy

0.90

 imply

0.89

 violate

0.89

Activations Density 0.616%

occurrences of the word "to" with various contexts and phrases

No Comments

No Known Activations