INDEX

Explanations

phrases indicating a cause or reason for a particular outcome

oai_token-act-pair · gpt-3.5-turbo

phrases indicating causality or consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 reinvent

-0.71

 quietly

-0.69

 sack

-0.67

 electr

-0.65

 tack

-0.63

 dwell

-0.62

 winning

-0.61

 sinking

-0.60

horn

-0.60

 sprint

-0.60

POSITIVE LOGITS

Due

3.36

Due

2.46

due

1.63

 Upon

1.29

 Depending

1.28

due

1.26

 Because

1.22

 Since

1.20

 Unfortunately

1.17

 Prior

1.16

Activations Density 0.015%

phrases indicating a cause or reason for a particular outcome

phrases indicating causality or consequences

No Comments

No Known Activations