INDEX

Explanations

proper nouns and names of people or entities

oai_token-act-pair · gpt-3.5-turbo

names of individuals involved in legal or political contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

eatures

-0.78

cellaneous

-0.68

</

-0.64

eaturing

-0.63

 Adds

-0.62

rawdownloadcloneembedreportprint

-0.61

ibilities

-0.61

 abound

-0.58

umph

-0.58

Characters

-0.58

POSITIVE LOGITS

 violated

1.48

 deserved

1.40

 misled

1.35

 interfered

1.29

 lacked

1.27

 acted

1.25

 misunderstood

1.24

 discriminated

1.24

 lied

1.23

 stole

1.21

Activations Density 0.355%

proper nouns and names of people or entities

names of individuals involved in legal or political contexts

No Comments

No Known Activations