INDEX

Explanations

phrases indicating a choice or preference

oai_token-act-pair · gpt-3.5-turbo

choices or decisions along with associated preferences

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

ankind

-0.76

ipples

-0.63

 latent

-0.58

ONY

-0.57

 disturbances

-0.56

 appre

-0.56

vind

-0.55

pires

-0.55

 glimps

-0.55

 incidents

-0.55

POSITIVE LOGITS

 instead

1.46

instead

1.22

 Instead

1.12

Instead

1.10

 because

1.05

 option

1.04

 rather

1.03

 alternatives

1.00

 lest

0.94

 anyways

0.92

Activations Density 0.565%

phrases indicating a choice or preference

choices or decisions along with associated preferences

No Comments

No Known Activations