INDEX

Explanations

locations or places described in a historical or news context

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

ãĥĦ

-0.81

ãĥ«

-0.76

oho

-0.69

pring

-0.69

ebook

-0.66

lets

-0.63

sth

-0.63

uner

-0.62

hift

-0.62

ãĥı

-0.61

POSITIVE LOGITS

 Postal

0.71

 Customs

0.70

 Superior

0.66

 Cellular

0.66

 Drops

0.65

 Virgin

0.63

icio

0.63

 Navy

0.63

 Torn

0.62

Activations Density 0.444%

locations or places described in a historical or news context

No Comments

No Known Activations