INDEX

Explanations

terms related to public affairs and policies

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

Scroll

-0.79

 Scroll

-0.76

uana

-0.69

RANT

-0.69

xual

-0.69

wered

-0.69

nesota

-0.68

kson

-0.67

vell

-0.67

nian

-0.66

POSITIVE LOGITS

 relations

1.09

 servants

1.07

izing

0.99

 sector

0.97

ised

0.93

 servant

0.90

 outcry

0.89

 opinion

0.89

izes

0.85

relations

0.84

Activations Density 2.708%

terms related to public affairs and policies

No Comments

No Known Activations