INDEX

Explanations

political terminology, including names of politicians and specific policy terms

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

Lear

-0.81

 pestic

-0.80

é¾įå

-0.78

âĵĺ

-0.74

FIG

-0.74

mbuds

-0.74

DragonMagazine

-0.73

umn

-0.73

æ©

-0.73

rown

-0.70

POSITIVE LOGITS

 perpetrated

0.72

 outwe

0.71

 rests

0.69

 itself

0.69

 workforce

0.68

 debacle

0.67

 Accountability

0.66

 backlog

0.63

 outweigh

0.63

 violates

0.63

Activations Density 0.292%

political terminology, including names of politicians and specific policy terms

No Comments

No Known Activations