INDEX

Explanations

words related to utility and energy services

oai_token-act-pair · gpt-3.5-turbo

references to utility services or companies

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

bring

-0.86

thus

-0.81

eworld

-0.80

ago

-0.79

usher

-0.75

oning

-0.74

ellen

-0.72

angling

-0.70

orne

-0.70

reads

-0.70

POSITIVE LOGITS

 utility

1.34

 utilities

1.08

 recip

0.85

 Utility

0.84

 nomine

0.84

ilities

0.83

 conduc

0.82

 glim

0.79

 Utilities

0.78

 veter

0.74

Activations Density 0.006%

words related to utility and energy services

references to utility services or companies

No Comments

No Known Activations