INDEX

Explanations

personal pronouns followed by emotional state words

oai_token-act-pair · gpt-3.5-turbo

expressions of ownership or personal perspective

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

iatrics

-0.70

elman

-0.66

llah

-0.65

Ã§a

-0.64

skinned

-0.62

piring

-0.61

wrap

-0.60

 href

-0.60

unci

-0.59

isk

-0.58

POSITIVE LOGITS

 detriment

1.43

 dismay

1.28

 satisfaction

1.18

 amaz

1.15

 surprise

1.14

 disadvantage

1.13

 liking

1.12

 aston

1.11

 credit

1.05

 discredit

1.05

Activations Density 0.063%

personal pronouns followed by emotional state words

expressions of ownership or personal perspective

No Comments

No Known Activations