INDEX

Explanations

references to time periods and statistics related to trends and events

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

umbn

-0.81

pired

-0.63

 oath

-0.63

pires

-0.62

inness

-0.59

atform

-0.56

ezvous

-0.56

Appearances

-0.56

DCS

-0.55

 hypocr

-0.55

POSITIVE LOGITS

 thanks

1.10

 compared

1.06

due

1.05

 owing

1.04

 because

1.01

thanks

0.93

because

0.89

 despite

0.88

 (âĪĴ

0.85

due

0.84

Activations Density 0.269%

references to time periods and statistics related to trends and events

No Comments

No Known Activations