INDEX

Explanations

phrases with the prefix "re-" indicating repetition or reversal of an action

oai_token-act-pair · gpt-3.5-turbo

words related to re-evaluation and reconsideration

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

 dying

-0.72

 deaths

-0.69

CPR

-0.67

 rebutt

-0.66

 opposition

-0.66

 concession

-0.65

 successors

-0.65

 infertility

-0.64

 disapproval

-0.64

 extra

-0.63

POSITIVE LOGITS

establish

1.38

evaluate

1.34

route

1.28

wrote

1.24

uph

1.22

insert

1.20

kind

1.19

orient

1.18

acqu

1.17

interpret

1.17

Activations Density 0.041%

phrases with the prefix "re-" indicating repetition or reversal of an action

words related to re-evaluation and reconsideration

No Comments

No Known Activations