INDEX

Explanations

statistics or numerical values in the 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Agent

-0.67

advert

-0.63

craft

-0.60

 Clash

-0.59

brand

-0.58

 Hawth

-0.57

Kl

-0.57

 Chimera

-0.56

 Hunt

-0.56

rag

-0.56

POSITIVE LOGITS

%-

0.85

Â½

0.85

 consecutive

0.84

âĢ²

0.83

Â¢

0.81

ÃĹ

0.80

 ILCS

0.74

yen

0.73

th

0.73

 percent

0.73

Activations Density 0.163%

statistics or numerical values in the text

No Comments

No Known Activations