INDEX

Explanations

past tense verbs that indicate an action performed by an individual

oai_token-act-pair · gpt-3.5-turbo

the word "had" indicating past experiences or actions

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

 VICE

-0.61

 Values

-0.58

Examples

-0.52

 Citizens

-0.51

 Example

-0.51

Change

-0.50

âĺħ

-0.49

TBA

-0.48

 guiName

-0.48

Adds

-0.48

POSITIVE LOGITS

had

2.68

 hadn

2.23

had

2.03

 lacked

1.88

 existed

1.70

Had

1.69

 seemed

1.65

 belonged

1.62

 knew

1.62

 remained

1.61

Activations Density 0.162%

past tense verbs that indicate an action performed by an individual

the word "had" indicating past experiences or actions

No Comments

No Known Activations