INDEX

Explanations

locations or events mentioned with a high emphasis

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

ratulations

-0.72

gans

-0.72

 benefited

-0.71

Pac

-0.71

aceutical

-0.69

shell

-0.65

hops

-0.64

HTTP

-0.64

fill

-0.62

archy

-0.62

POSITIVE LOGITS

 least

1.26

onement

1.13

 halftime

1.07

rium

1.01

yp

0.98

 dusk

0.94

 midnight

0.92

 noon

0.91

 midday

0.89

las

0.88

Activations Density 0.406%

locations or events mentioned with a high emphasis

No Comments

No Known Activations