INDEX

Explanations

proper nouns and personal names

oai_token-act-pair · gpt-3.5-turbo

specific proper nouns and brand names

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

 cost

-0.40

 examiner

-0.39

 ribbon

-0.39

 coordinate

-0.39

 date

-0.38

 designated

-0.38

 price

-0.38

 determining

-0.38

 quality

-0.37

 link

-0.37

POSITIVE LOGITS

irez

0.60

ada

0.60

atz

0.57

eta

0.55

ich

0.54

ares

0.54

asy

0.53

ades

0.53

ande

0.53

ak

0.53

Activations Density 2.528%

proper nouns and personal names

specific proper nouns and brand names

No Comments

No Known Activations