INDEX

Explanations

responses to official or important documents or conversational prompts

oai_token-act-pair · gpt-3.5-turbo

the character "ł" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

enegger

-0.81

 apes

-0.78

 condem

-0.78

raints

-0.74

 trainers

-0.65

 Seym

-0.65

 Instr

-0.65

 Journals

-0.65

 reflex

-0.64

 tort

-0.64

POSITIVE LOGITS

âĶĢâĶĢ

1.08

ï¸ı

1.05

ishable

1.00

uthor

0.99

vernment

0.95

ternity

0.89

nder

0.88

Â°

0.87

âĢł

0.87

âĶĢâĶĢâĶĢâĶĢ

0.86

Activations Density 0.191%

responses to official or important documents or conversational prompts

the character "ł" in various contexts

No Comments

No Known Activations