INDEX

Explanations

the word "Great" with a high activation strength

oai_token-act-pair · gpt-3.5-turbo

references to significant historical events, particularly those denoted by the term "Great."

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

quo

-0.73

qqa

-0.63

rett

-0.63

pport

-0.62

lessly

-0.61

mitt

-0.61

pta

-0.60

 dictated

-0.60

ded

-0.58

mitted

-0.58

POSITIVE LOGITS

 Recession

1.23

 Depression

1.09

 Lakes

1.04

 Barrier

0.98

 Plains

0.92

 Basin

0.91

 Divide

0.91

batch

0.89

 Pumpkin

0.86

 Dane

0.85

Activations Density 0.022%

the word "Great" with a high activation strength

references to significant historical events, particularly those denoted by the term "Great."

No Comments

No Known Activations