INDEX

Explanations

verbs related to actions or possibilities

oai_token-act-pair · gpt-3.5-turbo

concepts of ambiguity and generality

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

 guiActiveUnfocused

-0.80

ĸļ

-0.71

Ore

-0.64

Ranked

-0.62

asus

-0.61

HUD

-0.61

catentry

-0.60

govtrack

-0.60

 Then

-0.58

displayText

-0.58

POSITIVE LOGITS

 anywhere

1.64

 anytime

1.48

 anything

1.28

any

1.28

ANY

1.21

 either

1.20

 whichever

1.19

 whatever

1.13

 indefinitely

1.04

anything

1.03

Activations Density 0.455%

verbs related to actions or possibilities

concepts of ambiguity and generality

No Comments

No Known Activations