INDEX

Explanations

instructions or steps related to using a software tool or system

oai_token-act-pair · gpt-3.5-turbo

phrases related to the ability or capacity to perform actions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

breeding

-0.71

ylum

-0.68

 Mour

-0.66

 Hits

-0.66

 Notting

-0.61

 Seeking

-0.61

 Federation

-0.61

WOR

-0.60

polit

-0.60

ELD

-0.59

POSITIVE LOGITS

 optionally

1.43

't

1.39

 customize

1.25

 choose

1.24

 easily

1.22

 utilize

1.21

 alternatively

1.17

 specify

1.15

use

1.11

 modify

1.10

Activations Density 0.101%

instructions or steps related to using a software tool or system

phrases related to the ability or capacity to perform actions

No Comments

No Known Activations