INDEX

Explanations

phrases indicating a potential consequence or alternative option

oai_token-act-pair · gpt-3.5-turbo

conditional phrases that imply potential consequences or risks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

emouth

-0.71

onday

-0.66

owered

-0.66

ascus

-0.65

ENE

-0.64

essential

-0.63

vertising

-0.62

erning

-0.62

stic

-0.61

ãĤ·ãĥ£

-0.61

POSITIVE LOGITS

 else

1.61

Else

1.24

 alternatively

1.21

 perish

1.15

acles

1.13

chard

1.06

 starve

1.03

 risk

1.02

 succumb

0.97

acle

0.97

Activations Density 0.126%

phrases indicating a potential consequence or alternative option

conditional phrases that imply potential consequences or risks

No Comments

No Known Activations