INDEX

Explanations

contractions of "don't"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "don’t."

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

 ejected

-0.64

pus

-0.62

 protected

-0.61

 Goat

-0.60

 parting

-0.59

 Phase

-0.58

 Zoro

-0.57

 Published

-0.56

ipel

-0.56

 scrambling

-0.56

POSITIVE LOGITS

't

1.65

ned

1.24

uts

1.07

ates

1.05

eness

1.00

ÃŃ

0.94

ate

0.94

ated

0.92

hips

0.92

etsk

0.91

Activations Density 0.138%

contractions of "don't"

instances of the word "don’t."

No Comments

No Known Activations