INDEX

Explanations

phrases indicating conditions, reasoning, or causation

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

conditional phrases and concepts related to uncertainty or hypothetical scenarios

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 pione

-0.53

ãĥ¯ãĥ³

-0.49

erenn

-0.48

ãĤ¨ãĥ«

-0.47

eteen

-0.47

abre

-0.46

venth

-0.46

ãĤ´ãĥ³

-0.45

ãĤ©

-0.45

ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ

-0.44

POSITIVE LOGITS

she

0.66

SHE

0.61

he

0.61

 THEY

0.54

 they

0.52

0.51

 Xiaomi

0.49

we

0.49

 Kejriwal

0.48

you

0.48

Activations Density 1.697%

phrases indicating conditions, reasoning, or causation

conditional phrases and concepts related to uncertainty or hypothetical scenarios

No Comments

No Known Activations