INDEX

Explanations

phrases encouraging or instructing readers to continue engaging with content or take certain actions

oai_token-act-pair · gpt-3.5-turbo

the phrase "Keep" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

NetMessage

-0.68

 lowly

-0.67

Interstitial

-0.67

Jub

-0.64

ELF

-0.63

%);

-0.63

NRS

-0.62

é¾įåĸļå£«

-0.61

 crust

-0.61

 pupil

-0.61

POSITIVE LOGITS

 Keep

1.10

Keep

0.98

watch

0.91

keeping

0.83

Track

0.82

ers

0.80

warm

0.78

chin

0.77

keep

0.77

overs

0.75

Activations Density 0.009%

phrases encouraging or instructing readers to continue engaging with content or take certain actions

the phrase "Keep" in various contexts

No Comments

No Known Activations