INDEX

Explanations

places or things to which an action is being done

oai_token-act-pair · gpt-3.5-turbo

references to grouping or categorizing items or entities in a specific context

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

cest

-0.76

cial

-0.74

pmwiki

-0.74

lessness

-0.72

paio

-0.72

phal

-0.70

zees

-0.70

umar

-0.69

ociate

-0.67

trak

-0.67

POSITIVE LOGITS

 repertoire

1.52

 equation

1.43

 list

1.33

 arsenal

1.32

mix

1.29

 lineup

1.27

 reperto

1.14

 lists

1.03

 portfolio

1.02

 curriculum

1.01

Activations Density 0.245%

places or things to which an action is being done

references to grouping or categorizing items or entities in a specific context

No Comments

No Known Activations