INDEX

Explanations

phrases related to abstract concepts and philosophical discussions

oai_token-act-pair · gpt-3.5-turbo

unusual or non-standard characters and symbols

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

 favourite

-0.80

 favourites

-0.69

 Alpine

-0.67

PI

-0.65

 behavioural

-0.63

 defence

-0.61

QC

-0.59

PF

-0.58

PG

-0.57

QL

-0.57

POSITIVE LOGITS

 âĢķ

4.28

âĢķ

2.31

––

1.58

 âĶ

1.42

âĶĢ

1.38

âĶĢâĶĢ

1.26

—

1.25

--

1.18

ADVERTISEMENT

1.05

 âĶľ

1.04

Activations Density 0.008%

phrases related to abstract concepts and philosophical discussions

unusual or non-standard characters and symbols

No Comments

No Known Activations