INDEX

Explanations

phrases related to significant and wide-ranging impacts or consequences

oai_token-act-pair · gpt-3.5-turbo

phrases indicating extensive or significant consequences or effects

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

 Bulldogs

-0.66

Ŀ

-0.61

 Detect

-0.61

 Hyper

-0.60

 partName

-0.60

Hyper

-0.59

 certify

-0.58

VA

-0.58

%-

-0.57

aved

-0.57

POSITIVE LOGITS

reaching

1.13

ranging

0.96

issance

0.93

izophren

0.88

tenance

0.83

 philos

0.81

theless

0.80

emon

0.75

 compromises

0.74

redients

0.74

Activations Density 0.013%

phrases related to significant and wide-ranging impacts or consequences

phrases indicating extensive or significant consequences or effects

No Comments

No Known Activations