INDEX

Explanations

specific phrases or names

oai_token-act-pair · gpt-3.5-turbo

phrases related to formal titles or designations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

etheless

-0.56

RESULTS

-0.50

gomery

-0.49

 convincing

-0.46

 defamation

-0.44

 perjury

-0.44

NetMessage

-0.44

 quick

-0.44

Downloadha

-0.43

SPONSORED

-0.43

POSITIVE LOGITS

edin

0.61

 Rooms

0.60

 Room

0.58

 Centre

0.55

 Club

0.55

 Programme

0.55

 Alone

0.54

oran

0.54

 Point

0.53

 Thing

0.53

Activations Density 1.061%

specific phrases or names

phrases related to formal titles or designations

No Comments

No Known Activations