INDEX

Explanations

company names

oai_token-act-pair · gpt-3.5-turbo

prominent names and entities, particularly related to media, politics, and notable organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 fame

-0.63

tnc

-0.59

 behalf

-0.56

}.

-0.54

pron

-0.54

clave

-0.52

 guiActiveUnfocused

-0.52

EStreamFrame

-0.52

 vicinity

-0.50

LEASE

-0.49

POSITIVE LOGITS

 programmers

0.56

wat

0.55

 analysts

0.53

ensibly

0.52

 astronomers

0.50

 crept

0.50

 historians

0.49

 advocates

0.49

 commissions

0.48

 marketers

0.48

Activations Density 1.861%

company names

prominent names and entities, particularly related to media, politics, and notable organizations

No Comments

No Known Activations