INDEX

Explanations

phrases related to statements or quotes made by individuals

oai_token-act-pair · gpt-3.5-turbo

reported speech, particularly phrases indicating what someone said or wrote

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

 underestimated

-0.52

 perfected

-0.45

âī

-0.45

[+]

-0.45

ãĥİ

-0.44

 populous

-0.42

 Profession

-0.42

âĻ¦

-0.42

 Ranked

-0.41

iencies

-0.41

POSITIVE LOGITS

 sarcast

0.63

 referring

0.61

 incred

0.57

 rhet

0.57

 smir

0.57

 adds

0.56

 Blasio

0.56

 adding

0.56

 defiant

0.55

 Nasa

0.54

Activations Density 0.237%

phrases related to statements or quotes made by individuals

reported speech, particularly phrases indicating what someone said or wrote

No Comments

No Known Activations