INDEX

Explanations

phrases describing difficulty, challenges, or negative situations

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the potential outcomes or effects of actions

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

pta

-0.88

oday

-0.82

aucas

-0.74

eday

-0.73

phia

-0.70

orem

-0.69

psc

-0.68

terday

-0.68

 Previous

-0.68

orsi

-0.68

POSITIVE LOGITS

 frustrating

1.28

 daunting

1.28

 tricky

1.24

 intimidating

1.24

 devastating

1.19

 deadly

1.18

 debilitating

1.15

 confusing

1.15

 stressful

1.14

 troublesome

1.13

Activations Density 0.179%

phrases describing difficulty, challenges, or negative situations

phrases indicating the potential outcomes or effects of actions

No Comments

No Known Activations