INDEX

Explanations

phrases indicating actions or requests involving persuasion and influence

oai_token-act-pair · gpt-4o-mini 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

 Illustrated

-0.82

 largeDownload

-0.82

 Spoiler

-0.75

laughter

-0.71

 Printed

-0.69

 Definitions

-0.62

 Statistics

-0.62

DAQ

-0.61

Compar

-0.61

 Founding

-0.61

POSITIVE LOGITS

 accept

1.06

 obey

1.04

 cooperate

1.01

 behave

0.99

 surrender

0.98

 abandon

0.96

 undertake

0.95

 reconsider

0.93

 commit

0.93

 participate

0.92

Activations Density 0.180%

phrases indicating actions or requests involving persuasion and influence

No Comments

No Known Activations