INDEX

Explanations

questions or expressions of disbelief about situations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vity

-0.83

auri

-0.82

staking

-0.76

mur

-0.74

ouble

-0.74

inel

-0.72

onduct

-0.71

worn

-0.69

geoning

-0.69

avan

-0.68

POSITIVE LOGITS

 heck

1.49

 hell

1.29

 fuss

1.21

 fuck

1.06

 difference

0.83

 future

0.81

 average

0.79

 ramifications

0.79

 HELL

0.78

 Heck

0.77

Activations Density 0.065%

questions or expressions of disbelief about situations

No Comments

No Known Activations