INDEX

Explanations

pairs or groupings of related objects or concepts

oai_token-act-pair · gpt-3.5-turbo

highly descriptive nouns related to luxury items and significant cultural topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

putable

-0.74

Pg

-0.72

...]

-0.64

Dist

-0.63

given

-0.62

ollo

-0.61

felt

-0.61

arching

-0.61

src

-0.60

untled

-0.60

POSITIVE LOGITS

 gimmick

0.63

?????-

0.62

naire

0.61

©¶æ

0.60

ttes

0.58

 outwe

0.58

aments

0.57

 sandwic

0.57

 Mellon

0.57

MMO

0.55

Activations Density 0.725%

pairs or groupings of related objects or concepts

highly descriptive nouns related to luxury items and significant cultural topics

No Comments

No Known Activations