INDEX

Explanations

specific patterns like numbered lists and bullet points

oai_token-act-pair · gpt-3.5-turbo

structured information and procedural guidelines

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

iewicz

-0.71

bye

-0.66

mare

-0.66

 Strongh

-0.65

 unlaw

-0.64

 lore

-0.64

xus

-0.63

 Curious

-0.62

 neighb

-0.61

 scripting

-0.61

POSITIVE LOGITS

âĢ¢

0.95

âĹı

0.91

âĸł

0.86

Â·

0.80

audi

0.77

âĻ¦

0.74

Â¨

0.74

Whereas

0.71

ever

0.70

âľ

0.70

Activations Density 0.189%

specific patterns like numbered lists and bullet points

structured information and procedural guidelines

No Comments

No Known Activations