INDEX

Explanations

phrases related to groups of people or teams, especially referring to specific roles or organizations

oai_token-act-pair · gpt-3.5-turbo

terms related to groups, teams, or collections of people

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

urate

-0.66

ãĥĸ

-0.65

éģ

-0.63

actionDate

-0.63

 Plate

-0.62

ãģ

-0.61

 Related

-0.60

é¾įå¥ĳå£«

-0.57

FINE

-0.54

LEASE

-0.54

POSITIVE LOGITS

 consists

1.38

 consisted

1.35

 comprises

1.27

 includes

1.15

members

1.15

 gathers

1.13

 prefers

1.07

 welcomes

1.07

 disbanded

1.07

 expects

1.04

Activations Density 0.285%

phrases related to groups of people or teams, especially referring to specific roles or organizations

terms related to groups, teams, or collections of people

No Comments

No Known Activations