INDEX

Explanations

the word "ju" or words containing it

oai_token-act-pair · gpt-3.5-turbo

occurrences of the sequence "ju" in text, indicating a focus on a specific word or name

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Spectrum

-0.80

tons

-0.77

 Regulations

-0.75

 Orbit

-0.72

 Horizons

-0.71

space

-0.67

 Ninth

-0.67

 Advent

-0.67

graduate

-0.65

 Grain

-0.65

POSITIVE LOGITS

xtap

1.30

ju

1.02

ipers

1.00

arez

0.99

ice

0.98

iced

0.98

amph

0.96

udeau

0.96

pload

0.96

ulkan

0.92

Activations Density 0.010%

the word "ju" or words containing it

occurrences of the sequence "ju" in text, indicating a focus on a specific word or name

No Comments

No Known Activations