INDEX

Explanations

instances where someone is making a decision

oai_token-act-pair · gpt-3.5-turbo

instances of the word "to" as part of decision-making or intention statements

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

Lear

-0.74

ģĸ

-0.69

ĨĴ

-0.66

ĸļ

-0.65

ivia

-0.61

thumbnails

-0.59

grounds

-0.58

ãĤ¦ãĤ¹

-0.57

İ

-0.56

clips

-0.55

POSITIVE LOGITS

 pursue

1.22

 postpone

1.18

 ignore

1.13

 dedicate

1.08

 omit

1.08

 abandon

1.06

 skip

1.05

 disregard

1.04

 retire

1.04

 invest

1.03

Activations Density 0.179%

instances where someone is making a decision

instances of the word "to" as part of decision-making or intention statements

No Comments

No Known Activations