INDEX

Explanations

instances of verbs related to observation or perception

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

eware

-0.70

TO

-0.70

scribe

-0.68

ãĥĦ

-0.67

anium

-0.65

yright

-0.64

TL

-0.61

covered

-0.61

ÃŃa

-0.60

bes

-0.60

POSITIVE LOGITS

 firsthand

1.01

 glimps

0.93

 examples

0.91

 plenty

0.90

 parallels

0.84

 visions

0.83

 signs

0.82

 alot

0.80

 lots

0.80

dust

0.79

Activations Density 0.076%

instances of verbs related to observation or perception

No Comments

No Known Activations