INDEX

Explanations

personal pronouns and verbs related to giving instructions or suggestions

oai_token-act-pair · gpt-3.5-turbo

statements of certainty or predictions about future events

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

ablishment

-0.71

 Originally

-0.70

 disparate

-0.65

Gender

-0.60

 simultaneous

-0.57

 plurality

-0.57

 segregated

-0.57

Abstract

-0.57

 Stras

-0.57

Had

-0.57

POSITIVE LOGITS

 WILL

1.23

'll

1.15

 will

1.08

 wont

1.05

won

0.96

 might

0.91

may

0.90

may

0.89

MAY

0.88

 SHALL

0.87

Activations Density 0.374%

personal pronouns and verbs related to giving instructions or suggestions

statements of certainty or predictions about future events

No Comments

No Known Activations