INDEX

Explanations

tips or suggestions related to tools, changes, or solutions

oai_token-act-pair · gpt-3.5-turbo

keywords related to various forms of support and resources

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 seiz

-0.57

emale

-0.57

 Rossi

-0.56

 palp

-0.55

phas

-0.55

tesque

-0.53

theless

-0.53

 Gleaming

-0.53

 Contrast

-0.51

lishes

-0.51

POSITIVE LOGITS

âĢ

1.22

*.

0.95

.*

0.94

âĿ

0.94

.ï¿½

0.92

.</

0.91

 someday

0.86

 âĿ

0.85

 ðŁĺ

0.84

âľ

0.84

Activations Density 0.671%

tips or suggestions related to tools, changes, or solutions

keywords related to various forms of support and resources

No Comments

No Known Activations