INDEX

Explanations

numerical values or product codes

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

yip

-0.69

 reckoning

-0.63

boxing

-0.62

abouts

-0.61

fork

-0.59

EStream

-0.58

byn

-0.58

Ô

-0.57

 Alban

-0.56

 è£ıè

-0.56

POSITIVE LOGITS

ciating

1.19

iencies

1.17

ments

1.09

ment

1.07

itely

1.05

uration

1.05

aneous

1.05

entials

1.04

ption

1.04

ential

1.04

Activations Density 1.011%

numerical values or product codes

No Comments

No Known Activations