INDEX

Explanations

the word "don't."

oai_token-act-pair · gpt-3.5-turbo

negative contractions of "do not"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

soType

-0.70

itiz

-0.69

Reviewer

-0.67

 Reloaded

-0.67

 Penguin

-0.64

estern

-0.64

çĦ

-0.64

edIn

-0.63

Gry

-0.59

pter

-0.59

POSITIVE LOGITS

 necessarily

1.16

 bother

1.04

 know

0.86

 anymore

0.86

 seem

0.86

necess

0.86

 intend

0.85

urtles

0.85

 appreciate

0.85

 expect

0.84

Activations Density 0.104%

the word "don't."

negative contractions of "do not"

No Comments

No Known Activations