INDEX

Explanations

contractions of "do not" or "does not"

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 nearest

-0.69

 enthusi

-0.67

 populated

-0.65

 newcom

-0.64

 redes

-0.63

 cancellation

-0.63

 Cance

-0.62

 exha

-0.61

 Published

-0.61

 princ

-0.61

POSITIVE LOGITS

't

1.84

ÃŃ

1.09

uts

1.00

eness

0.96

Â´

0.94

etsk

0.92

ned

0.92

hips

0.87

ALD

0.84

Activations Density 0.153%

contractions of "do not" or "does not"

No Comments

No Known Activations