INDEX

Explanations

words related to positioning or attaching something

oai_token-act-pair · gpt-3.5-turbo

actions related to fastening, attaching, or integrating items

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

yan

-0.69

DOM

-0.63

QUI

-0.62

taker

-0.59

Ĥİ

-0.59

edom

-0.58

sein

-0.58

stal

-0.58

Represent

-0.58

 tending

-0.57

POSITIVE LOGITS

 onto

0.95

 Into

0.87

 into

0.86

 INTO

0.79

 together

0.77

 overboard

0.75

innie

0.73

 dots

0.70

 elements

0.69

endum

0.68

Activations Density 0.271%

words related to positioning or attaching something

actions related to fastening, attaching, or integrating items

No Comments

No Known Activations