INDEX

Explanations

words related to forceful actions or impacts

oai_token-act-pair · gpt-3.5-turbo

words related to dramatic or impactful actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PF

-0.78

DOC

-0.70

ob

-0.67

 broch

-0.63

 science

-0.63

Ik

-0.63

 Norway

-0.62

 Frey

-0.62

 ther

-0.61

 amen

-0.61

POSITIVE LOGITS

ashing

4.17

ashed

2.79

ashes

2.70

ASH

2.03

ash

1.79

asher

1.35

attering

1.30

atching

1.25

usting

1.22

acking

1.11

Activations Density 0.005%

words related to forceful actions or impacts

words related to dramatic or impactful actions

No Comments

No Known Activations