INDEX

Explanations

expressions related to discussions or reflections on various topics

oai_token-act-pair · gpt-3.5-turbo

expressions of hesitation, caution, or shame regarding personal experiences or opinions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ngth

-0.80

ynthesis

-0.79

strous

-0.73

vantage

-0.71

ictionary

-0.66

DragonMagazine

-0.66

anooga

-0.65

odder

-0.65

orld

-0.64

uction

-0.64

POSITIVE LOGITS

how

1.21

 what

1.05

 admitting

1.04

 acknowledging

1.02

 whether

1.01

 letting

0.99

 where

0.96

 choosing

0.95

 disclosing

0.94

 knowing

0.94

Activations Density 0.187%

expressions related to discussions or reflections on various topics

expressions of hesitation, caution, or shame regarding personal experiences or opinions

No Comments

No Known Activations