INDEX

Explanations

articles ('a', 'an') followed by a noun

oai_token-act-pair · gpt-3.5-turbo

articles and their usage in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

grounds

-0.74

 except

-0.72

 reacts

-0.72

tarians

-0.70

 adherents

-0.69

aos

-0.69

 shares

-0.69

 produces

-0.68

Examples

-0.67

 instruments

-0.67

POSITIVE LOGITS

 longstanding

1.03

 handful

1.02

 slew

1.01

lot

0.97

 dozen

0.96

 plethora

0.95

 whopping

0.95

 multitude

0.94

 swath

0.94

few

0.94

Activations Density 0.342%

articles ('a', 'an') followed by a noun

articles and their usage in context

No Comments

No Known Activations