INDEX

Explanations

names of people, locations, and medical terms

oai_token-act-pair · gpt-3.5-turbo

names and terms associated with the Hakusho franchise

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Papers

-0.84

 Survivors

-0.80

lihood

-0.75

 Pompe

-0.75

 Commonwealth

-0.73

 Staples

-0.66

 Sterling

-0.66

 Pearce

-0.66

 Patton

-0.65

ãģ®éŃĶ

-0.64

POSITIVE LOGITS

Hak

1.01

uno

0.92

mares

0.85

stall

0.84

hig

0.84

KEN

0.83

odi

0.82

mare

0.81

bang

0.80

lass

0.80

Activations Density 0.025%

names of people, locations, and medical terms

names and terms associated with the Hakusho franchise

No Comments

No Known Activations