INDEX

Explanations

verbs related to observation and perception

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate comprehension or observation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

ranged

-0.66

uga

-0.66

eware

-0.63

etheless

-0.62

nation

-0.62

inqu

-0.62

phant

-0.61

ever

-0.60

icum

-0.59

upe

-0.59

POSITIVE LOGITS

why

1.39

how

1.15

WHY

1.00

 whats

1.00

why

0.99

 clearly

0.92

 whence

0.88

 glimps

0.88

 traces

0.87

 similarities

0.86

Activations Density 0.086%

verbs related to observation and perception

phrases that indicate comprehension or observation

No Comments

No Known Activations