INDEX

Explanations

numerical values indicating a minimum threshold

oai_token-act-pair · gpt-3.5-turbo

phrases indicating minimum thresholds or requirements

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

eur

-0.60

REP

-0.58

auer

-0.57

Sense

-0.54

rocket

-0.53

john

-0.53

hunter

-0.52

AMA

-0.52

 respectively

-0.52

CRIP

-0.51

POSITIVE LOGITS

 least

3.55

 slightest

1.75

 smallest

1.43

 weakest

1.38

 shortest

1.35

 lowest

1.30

 cheapest

1.25

 widest

1.24

 longest

1.22

 safest

1.18

Activations Density 0.034%

numerical values indicating a minimum threshold

phrases indicating minimum thresholds or requirements

No Comments

No Known Activations