INDEX

Explanations

words related to a specific topic or organization, possibly with a focus on insider information

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "Ins" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

âĢ¢âĢ¢

-0.79

=-=-=-=-=-=-=-=-

-0.76

BY

-0.72

lda

-0.71

âĺħâĺħ

-0.71

ä»£

-0.71

AUT

-0.70

DragonMagazine

-0.70

=-=-=-=-

-0.70

 Carbuncle

-0.69

POSITIVE LOGITS

ufficient

1.13

urrection

1.03

iders

0.94

ights

0.93

eman

0.89

ensibly

0.87

urance

0.87

lamm

0.87

urers

0.85

cept

0.84

Activations Density 0.008%

words related to a specific topic or organization, possibly with a focus on insider information

occurrences of the term "Ins" in various contexts

No Comments

No Known Activations