INDEX

Explanations

the word "if"

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

ãĤª

-0.79

Eye

-0.71

akia

-0.65

iliar

-0.65

ãĥŃ

-0.64

 Nationwide

-0.63

oya

-0.60

FTWARE

-0.60

 Deity

-0.60

holm

-0.60

POSITIVE LOGITS

fy

1.14

rame

1.06

you

0.96

ornia

0.94

 anything

0.86

 they

0.82

 necessary

0.80

 unchecked

0.80

tar

0.78

 anybody

0.77

Activations Density 0.407%

the word "if"

No Comments

No Known Activations