INDEX

Explanations

questions or statements implying an action or decision

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

 faint

-0.69

inently

-0.67

rid

-0.65

printed

-0.64

 serv

-0.61

tri

-0.60

DOT

-0.60

rin

-0.60

 fisher

-0.59

gart

-0.58

POSITIVE LOGITS

 regarding

0.86

 next

0.75

 beforehand

0.74

 NEXT

0.73

?:

0.72

 tonight

0.71

 during

0.71

ãģĵ

0.68

 besides

0.67

 today

0.67

Activations Density 0.960%

questions or statements implying an action or decision

No Comments

No Known Activations