INDEX

Explanations

instances of the word "non" followed by a variety of different subsequent words or phrases

oai_token-act-pair · gpt-3.5-turbo

references to non-profit organizations and concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Tycoon

-0.97

 Orchestra

-0.82

 Gardens

-0.81

 Cobra

-0.79

 Spoon

-0.77

 Grind

-0.75

 Oaks

-0.74

 Sparrow

-0.74

 McGee

-0.74

 Memories

-0.73

POSITIVE LOGITS

etheless

1.02

fiction

1.01

zero

0.98

chal

0.96

basic

0.95

combat

0.91

epad

0.91

negative

0.91

availability

0.90

linear

0.90

Activations Density 0.025%

instances of the word "non" followed by a variety of different subsequent words or phrases

references to non-profit organizations and concepts

No Comments

No Known Activations