INDEX

Explanations

mentions of events, statistics, or updates in various locations

oai_token-act-pair · gpt-3.5-turbo

phrases that denote the existence or occurrence of events or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

tyr

-0.61

IPM

-0.60

rients

-0.59

 Ashes

-0.58

 Civilization

-0.56

 Dise

-0.55

 entr

-0.54

 Coffin

-0.54

 Slayer

-0.53

 Doom

-0.53

POSITIVE LOGITS

abouts

1.21

 were

1.16

after

1.09

are

1.08

 have

1.00

upon

0.99

 appears

0.98

 hasn

0.97

 weren

0.97

was

0.96

Activations Density 0.105%

mentions of events, statistics, or updates in various locations

phrases that denote the existence or occurrence of events or entities

No Comments

No Known Activations