INDEX

Explanations

words related to electrical items or devices

oai_token-act-pair · gpt-3.5-turbo

instances of the substring "ind."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 elig

-0.65

iannopoulos

-0.62

TI

-0.61

itton

-0.60

 advancements

-0.60

 stren

-0.59

 blazing

-0.58

 disbel

-0.58

 tremend

-0.56

©¶æ

-0.56

POSITIVE LOGITS

erella

1.24

ebted

1.21

icator

1.17

icative

1.09

icators

1.05

inav

1.00

ication

0.98

rift

0.96

say

0.96

irect

0.95

Activations Density 0.021%

words related to electrical items or devices

instances of the substring "ind."

No Comments

No Known Activations