INDEX

Explanations

first person pronouns

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

PTS

-0.73

 Plaint

-0.62

tnc

-0.59

 totality

-0.58

 Malone

-0.57

 relevance

-0.57

pires

-0.56

Ele

-0.56

 Pearson

-0.55

 antid

-0.55

POSITIVE LOGITS

'm

1.60

've

1.34

 dunno

1.32

 suppose

1.29

 guess

1.26

'll

1.26

 swear

1.16

'd

1.14

am

1.10

 wanna

1.07

Activations Density 0.204%

first person pronouns

No Comments

No Known Activations