INDEX

Explanations

phrases related to online activities and discussions

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

 Pearce

-0.73

ATURE

-0.72

 Stevens

-0.69

Ģ

-0.68

========

-0.67

ELS

-0.67

baugh

-0.67

Vaj

-0.65

maid

-0.65

eters

-0.64

POSITIVE LOGITS

 cafÃ©

0.87

 cafes

0.86

ibaba

0.83

 encyclopedia

0.82

 forums

0.82

caf

0.81

 retailer

0.81

 platforms

0.79

 cafe

0.78

 savvy

0.78

Activations Density 0.466%

phrases related to online activities and discussions

No Comments

No Known Activations