INDEX

Explanations

words related to a specific technology or company acronym

oai_token-act-pair · gpt-3.5-turbo

mentions of "SE" and associated programming or systems-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

iques

-0.83

itates

-0.82

iary

-0.81

ique

-0.80

itate

-0.79

idae

-0.79

selling

-0.78

icity

-0.77

istics

-0.77

itating

-0.76

POSITIVE LOGITS

VER

1.23

VEN

1.19

VEL

1.16

VE

1.14

IU

1.14

LECT

1.12

BOOK

1.12

VIS

1.12

ARCH

1.11

ASON

1.10

Activations Density 0.064%

words related to a specific technology or company acronym

mentions of "SE" and associated programming or systems-related terms

No Comments

No Known Activations