INDEX

Explanations

news-related terms and information

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

bun

-0.68

bat

-0.64

acia

-0.63

bern

-0.60

nib

-0.59

abba

-0.57

umble

-0.56

icum

-0.56

 charm

-0.56

formation

-0.55

POSITIVE LOGITS

 WATCH

0.81

 Watching

0.71

esome

0.70

 VIDEOS

0.69

 Sending

0.65

dog

0.64

...]

0.60

 Allows

0.60

------------------------------------------------

0.59

 View

0.59

Activations Density 0.006%

news-related terms and information

No Comments

No Known Activations