INDEX

Explanations

instances of the word "engaged" in the text

oai_token-act-pair · gpt-3.5-turbo

instances of the word "engaged" 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 @ 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rule

-0.76

 appearances

-0.63

 Donation

-0.62

ppings

-0.62

 gloom

-0.61

Zot

-0.60

grain

-0.60

 Rays

-0.59

 Curse

-0.59

Nug

-0.59

POSITIVE LOGITS

 engaged

3.57

 engage

2.09

 engaging

2.08

 engages

2.06

 engagement

1.76

 diseng

1.59

 involved

1.44

 committed

1.42

 immersed

1.41

 employed

1.40

Activations Density 0.019%

instances of the word "engaged" in the text

instances of the word "engaged" in various contexts

No Comments

No Known Activations