INDEX

Explanations

names of individuals

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

 gust

-0.29

 toppled

-0.29

 inflamm

-0.28

 volatile

-0.28

 fragmented

-0.28

 erratic

-0.27

 populism

-0.26

 lofty

-0.26

 populist

-0.26

 peripher

-0.26

POSITIVE LOGITS

POST

0.37

 =================================================================

0.36

Wiki

0.32

Sex

0.31

Coin

0.29

!!!

0.29

 ÂłÂł

0.29

HT

0.29

EH

0.29

combe

0.29

Activations Density 10.711%

names of individuals

No Comments

No Known Activations