INDEX

Explanations

phrases related to uncertainty or speculation

oai_token-act-pair · gpt-3.5-turbo

possibilities and speculations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

 rall

-0.66

noticed

-0.66

Awesome

-0.64

OV

-0.61

 Gets

-0.61

DAQ

-0.61

ashes

-0.61

ACP

-0.59

efeated

-0.59

verages

-0.58

POSITIVE LOGITS

 explain

1.28

 signify

1.26

 indicate

1.17

 represent

1.13

 imply

1.12

 reflect

1.12

 originate

1.11

be

1.09

 relate

1.09

 contribute

1.06

Activations Density 0.149%

phrases related to uncertainty or speculation

possibilities and speculations

No Comments

No Known Activations