INDEX

Explanations

phrases or names related to debates

oai_token-act-pair · gpt-3.5-turbo

references to debates and discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ĸļ

-0.75

ISTER

-0.71

ourke

-0.66

hetical

-0.65

ovie

-0.65

chell

-0.62

aer

-0.61

 actionGroup

-0.61

face

-0.61

imates

-0.60

POSITIVE LOGITS

deb

1.06

Deb

0.90

 mathemat

0.79

bie

0.77

 juven

0.76

 distingu

0.75

 resil

0.75

bles

0.75

ilitation

0.72

 contrace

0.72

Activations Density 0.005%

phrases or names related to debates

references to debates and discussions

No Comments

No Known Activations