INDEX

Explanations

phrases or words associated with giving positive feedback or admiration

oai_token-act-pair · gpt-3.5-turbo

contexts of admiration and commendation

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

arin

-0.75

gamer

-0.72

olen

-0.69

Â¯Â¯

-0.68

carry

-0.66

perature

-0.64

Hack

-0.61

Gamer

-0.60

otype

-0.60

NetMessage

-0.60

POSITIVE LOGITS

 praise

1.10

 praises

0.95

 praising

0.92

hovah

0.92

 praised

0.84

eous

0.83

ifully

0.80

 thanking

0.78

 acclaim

0.73

imaru

0.72

Activations Density 0.020%

phrases or words associated with giving positive feedback or admiration

contexts of admiration and commendation

No Comments

No Known Activations