INDEX

Explanations

phrases with instructions or requests

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

Bridge

-0.69

Hum

-0.67

osa

-0.65

EVA

-0.65

avery

-0.64

Â¨

-0.62

apo

-0.61

Prosecut

-0.60

forth

-0.57

ede

-0.57

POSITIVE LOGITS

 been

0.79

been

0.77

 subscribed

0.75

 ever

0.73

 trouble

0.72

 invested

0.71

 doubts

0.71

any

0.70

 bothered

0.68

 chosen

0.67

Activations Density 0.094%

phrases with instructions or requests

No Comments

No Known Activations