INDEX

Explanations

direct instructions or requests in text

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ö¼

-0.80

GoldMagikarp

-0.69

Downloadha

-0.66

Hop

-0.64

Traditional

-0.62

olver

-0.61

Pred

-0.61

MpServer

-0.58

Drive

-0.57

ELD

-0.57

POSITIVE LOGITS

 responsibly

0.85

 yourselves

0.81

 reprint

0.80

ruary

0.76

 PLEASE

0.74

 yourself

0.70

edia

0.69

 acknow

0.66

clus

0.65

estine

0.64

Activations Density 0.070%

direct instructions or requests in text

No Comments

No Known Activations